Research Colloquium

CICMHE sponsors the biannual International Material Handling Research Colloquium (IMHRC) to bring together academic and industrial participants to discuss the latest in material handling, facility layout and design, logistics and supply chain research. The Colloquium is held every two years in the summer of even-numbered years. The first IMHRC was held in 1990.

The research issues presented and discussed at each Colloquium are documented in a series of hard-bound edited books, entitled Progress in Material Handling Research. Past volumes are available through the MHI publications.

14th International Material Handling Research Colloquium – 2016
Progress in Material Handling Research: 2016

2016 IMHRC Table of Contents and Abstracts

Performance Comparison of Automated Warehouses Using Simulation
Nand Kishore Agrawal, Sunderesh S. Heragu and Chinnatat Methapatra The purpose of this study is to compare the performance of two types of warehouses, both of which use autonomous vehicles (AVs). One warehouse uses movable racks (MR) for storing mini-loads, whereas the other uses fixed racks (FR). In general, warehouse automation not only increases the speed of the fulfillment process but also makes the picking process more accurate. We simulate three scenarios for the MR and FR systems using Simio. Four performance measures are considered for the comparison – the average order processing time (WR), the average utilization of AVs (U), the average order processing queue length (Nq) and the average distance travelled by AVs (d). We also estimate the capital costs of both systems and use it to compare the two systems. On the basis of our assumptions and simulation results, we find that the FR system not only requires an average 56 % less capital investment than the MR system, but it also provides a more efficient warehousing automation option with relatively lower utilization of AVs, lower order processing time and lower average number of orders waiting to be processed.
Job Sequencing in a Miniload System
Lennart Baardman, Kees Jan Roodbergen and Héctor J.Carlo More and more warehouses shift towards the use of automated systems. One such system is a miniload system. In this system, a crane retrieves bins from their storage locations and brings the bins to the picking station, where a worker takes the requested products from the retrieved bins. After a product has been picked, the bin is put back by the crane in its original location. A buffer is present at the picking station to absorb fluctuations in speed between the picker and the crane. The problem of scheduling jobs in this system has received little attention in the literature. We describe system properties and give insight in the performance of a number of heuristics.
Transport Items and Physical Internet Handling Boxes: A Comparison Framework Across Supply Chains
Eric Ballot and Benoit Montreuil Pallets, cardboard boxes, and plastic crates are widely used tools to operate supply chains. As such they have many impacts on handling effort, shipment protection, transport mean utilization as well as repositioning and recycling efforts and they represent billions of assets spread all over the world. The historical and local origins of the designs and the sharing among many stakeholders do not ensure at all any kind of global optimization. The purpose of this paper is to define and explain a research effort to better measure and evaluate the efficiencies and inefficiencies for supply chain stakeholders for themselves and globally. When validated the framework will be sued to evaluate new designs on a global scale especially for the design of the handling box related to the Physical Internet concept.
Dual-Tray Vertical Lift Modules for Fast Order Picking
Daria Battini, Martina Calzavara, Alessandro Persona and Fabio Sgarbossa In the last years, new solutions for order picking systems have been developed both from industry and academics, especially for small items. They include innovative flexible automatic parts-to-picker systems and optimized picker-to-parts ones. One of these solutions consists in the use of Vertical Lift Modules (VLMs), a storage column in which small items are stored in extractable trays. In this paper, we study a new system composed by dual-tray VLMs where the operators perform picking and sorting activities. We propose several actions in order to improve the productivity of the entire system: 1) class based storage assignment of items inside the VLMs; 2) batch retrievals of items and 3) batch orders and batch retrievals with pick-and-sort activity. The impacts of these actions are evaluated with a simulation of the system using real data from an industrial case.
Continuous Approximation of Multi Cycle Time for Multi Aisles Automated Storage and Retrieval Systems
Yassine Boudghene Stambouli and Latefa Ghomri Knowing that, in an automated storage and retrieval system, the travel time of the storage/retrieval machine is an important parameter which affects the whole system performances, several works were dedicated to its evaluation and its modeling, in single, dual and multi command. These models were proposed for different configurations of AS/RS; such as unit load, multi-aisles and Mobile Racks. These models are based mainly on a discrete approach and the analytical expressions represent the system functioning exactly. In this paper we present a general approach for analytical modeling of multi command cycle time by giving particularly the modeling of the Time Between that we will use to model the Dual Cycle time and the Multi Cycle time. This continuous approach with its simplified mathematical expressions aims to make the calculations more easy than the discrete approach and can be generalized for any one deep physical configuration of AS/RSs and later used to make an optimization of the dimensions of such systems.
Throughput Rate of a Two-Worker Stochastic Bucket Brigade
Yossi Bukchin, Eran Hanany and Eugene Khmelnitsky Work sharing in production systems is a modern approach that improves throughput rate. Work is shifted between cross-trained workers in order to better balance the material flow in the system. When a serial system is concerned, a common work-sharing approach is the Bucket-Brigade (BB), by which downstream workers sequentially take over items from adjacent upstream workers. When the workers are located from slowest-to-fastest and their speeds are deterministic, it is known that the line does not suffer from blockage or starvation, and achieves the maximal theoretical throughput rate (TR). Very little is known in the literature on stochastic self-balancing systems with work sharing, and on BB in particular. This paper studies the basic BB model of Bartholdi & Eisenstein (1996) under the assumption of stochastic worker speeds. We identify settings in which conclusions that emerge from deterministic analysis fail to hold when speeds are stochastic, in particular relating to worker order assignment as a function of the problem parameters.
A Performance Measurement Framework and Solution Approach for the Integrated Facility Layout Problem with Uncertain Demand
Melih Çelik, Begün Efeoglu and Haldun Süral The integrated facility layout problem (IFLP) focuses on the simultaneous determination of the relative locations of multiple copies of capacitated equipment or machinery in a facility, as well as the material flow between these units. In this paper, we consider the IFLP in the existence of uncertain demand for the products of the facility. Motivated by the framework for next generation facility layouts by Benjaafar et al. (2002), we extend the approaches in the literature for distributed facility layouts to the case of dynamic demand and the possibility of relay outs, and propose a heuristic solution approach to minimize the expected total material handling cost over the planning horizon. We also analyze the performance of the resulting solutions in terms of empty travel of the material handling equipment and waiting time. Our computational results reveal that when demand is dynamic and stochastic, the relationship between the level of uncertainty and relay out cost plays an important role in determining layout performance, and therefore a priori assumption of using a certain layout type may lead to detrimental results.
A Simulation Model to Evaluate the Layout for Block Stacking Warehouses
Shahab Derhami, Jeffrey S. Smith and Kevin R. Gue Storing pallets of Stock Keeping Units (SKUs) on top of one another in lanes on a warehouse floor is known as block stacking. This storage system is widely used in manufacturing systems and distribution centers. The arrangement of lanes in the layout of this system significantly impacts utilization of the storage space and transportation costs. Existing research that studies the layout for this system focuses exclusively on determining the optimal lane depth with respect to space utilization and ignores transportation costs. In this study, we develop a simulation model that computes several performance metrics to evaluate both of these objectives for a warehouse layout. It aims to take the stochastic variations exist in the real world situation into account. Designing the layout based on the historical data distinguishes this model from the analytical models in the systems with high level of uncertainty, where determining the required parameters for analytical models are difficult due to the high variations. We verified the model using the existing analytical models and developed an experimental analysis to show the trade-off between the space utilization and transportation costs in the layout design problem.
Determination of Cycle Times for Double Deep Storage Systems Using a Dual Capacity Handling Device
Katharina Dörr and Kai Furmans Double deep storage is an efficient method to improve space utilization in warehouses. Contrary to intuition, it also can be efficient when considering retrieval times, since the aisles of a warehouse with double deep storage may be shorter than comparable warehouses with single deep storage. AS/RSmachines equipped with two load-handling units might further improve the situation, since two storage units (for instance pallets or cases) can be stored and retrieved, effectively allowing quadruple command cycles, bringing in total two storage units into the aisle and retrieving two storage units at the same time. In this paper, we present a method for the computation of cycle times and average cycle times with the assumption of equally distributed access probabilities for AS/RS-machines equipped with two load handling devices.
Energy and Cycle Time Efficient Warehouse Design for Autonomous Vehicle-Based Storage and Retrieval System
Banu Y. Ekren, Anil Akpunar and Tone Lerher This study explores the best warehouse design for an autonomous vehicle based storage and retrieval system (AVS/RS) minimizing average energy consumption per transaction and average cycle time per transaction, simultaneously. In the design concept, we consider, rack design in terms of number of bays, number of tiers, number of aisles; number of resources, namely number of autonomous vehicles and lifts and; velocity profiles of lifts and autonomous vehicles in the AVS/RS. We completed 1,296 number of experiments in simulation to obtain Pareto solutions representing the “average energy consumption per transaction” and “average cycle time per transaction” trade-offs based on designs which is a very useful visual tool in decision making. Different from the existing studies, we approach to the warehouse design problem of AVS/RSs from a multi-objective view as well as energy efficient view minimizing both electricity consumption and cycle time per transaction in the system.
A Scalable Algorithm for Locating Distribution Centers on Real Road Networks
Saeed Ghanbartehrani and J. David Porter The median problem is a type of network location problem that aims at finding a node with the total minimum demand weighted distance to a set of demand nodes in a weighted graph. In this research, an algorithm for solving the median problem on real road networks is proposed. The proposed algorithm, referred to as the multi-threaded Dijkstra’s (MTD) algorithm, is then used to optimally locate Wal-Mart distribution centers on the 28-million node road network of the United States with the objective of minimizing the total demand weighted transportation cost. The resulting optimal location configuration of Wal-Mart distribution centers improves the total transportation cost by 40%.
A High-Density, Puzzle-Based System for Rail-Rail Container Transfers
Kevin R. Gue and Gang Hao We describe a high-density, puzzle-based storage and transfer system for containers in a rail-torail hub for the Physical Internet. The system uses a new algorithm called GridHub, which is able to transfer items in all four cardinal directions simultaneously within a grid. We show how the GridHub system might be used in a rail-rail transfer hub to transfer containers between one side of the grid and a train.
A Novel Approach to Analyze Inventory Allocation Decisions in Robotic Mobile Fulfillment Systems
T. Lamballais, D. Roy and M.B.M. de Koster The Robotic Mobile Fulfillment System is a newly developed automated, parts-to-picker material handling system. Storage shelves, also known as inventory pods, are moved by robots between the storage area and the workstations, which means that they can be continually repositioned during operations. This paper develops a queueing model for optimizing three key decision variables: (1) the number of pods per product (2) the ratio of the number of pick to the number of replenishment stations, and (3) the replenishment level per pod. We show that too few or too many pods per product leads to unnecessarily long order throughput times, that the ratio of the number of pick to the number of replenishment stations can be optimized for order throughput time, and that waiting to replenish until a pod is completely empty can severely decrease throughput performance.
Simulation-Based Energy and Cycle Time Analysis of Shuttle-Based Storage and Retrieval System
Tone Lerher, Banu Y. Ekren and Anil Akpunar This study explores the best warehouse design for shuttle-based storage and retrieval system (SBS/RS) minimizing average energy consumption per transaction and average cycle time per transaction, simultaneously. For that we provided average energy consumption per transaction versus average cycle time per transaction graphs, for different design scenarios of the studied SBS/RS warehouse. In the design concept, we considered, rack design in terms of number of bays, number of tiers, number of aisles, as well as velocity profiles of lifts in the system. We completed 144 experiments by simulation to see the trade-offs based on the design scenarios and provided them by two separate graphs. The results show that while the SBS/RS warehouse has low number of tiers, it has low energy consumption per transaction as well as low average cycle time per transaction in the two lift velocity scenarios.
Introducing the Concept of Hyperconnected Mobile Production
Suzanne Marcotte and Benoit Montreuil Many globalized businesses are trying to cope with growing competition by strategically expanding their dedicated network of production facilities so as to be able to offer and to deliver time and price competitive offers to their clients across the world. In this paper, exploiting Physical Internet principles, we introduce the concept of hyperconnected mobile production that can alternatively enable businesses to dynamically expand and contract as necessary their production capacity in regions worldwide. First, hyperconnected mobile production exploits open fabs from multiple parties readily available in those regions. Second, these fabs are to rely on plug-and-play production modules. These modules are to be flowed in and out of open fabs worldwide by the fab operators or their business clients so as to absorb dynamic production requirements from customers. Third, the production modules are to be dynamically reconfigurable through adding and removing plug-and-play modular resources. We first show that hyperconnected mobile production builds on eight innovation threads: distributed, outsourced, on-demand, modular, additive, mobile, containerized and hyperconnected production. We then provide an overall description of key facets of the hyperconnected mobile production concept and finally elicit a number of promising research avenues.
Operational-Level Optimization of Inbound Intralogistics
Yeiram Martínez and Héctor J. Carlo This study is concerned with optimizing inbound operations at distribution centers (DCs), warehouses, and cross-docks with staging areas. The objective of the problem is to minimize the makespan required to move all unit loads from the trailers to the flow racks, and from the flow racks to their respective storage locations. It is assumed that a set of inbound trailers with known composition have been assigned and sequenced to inbound dock doors. The following three inbound logistics decisions are simultaneously considered: i) unloaders’ assignment and scheduling, ii) loads-to-flow rack assignment, and iii) assignment and haulers’ scheduling. In this study we describe the relationship between the problem of minimizing makespan and an unloader-hauler balancing problem. Three rule-based heuristics are proposed and evaluated in an instance of the problem.
Non-Traditional Aisle Design for a Manufacturing Facility Layout
Dale Masel and Dean Marinchek Methods for designing a facility layout typically assume that the aisles for interdepartmental travel will be parallel to the exterior walls of the facility. However, by putting aisles at an angle to the exterior walls, more direct routes between departments can be created, reducing travel distance. This paper describes a method to create a facility layout that does not have orthogonal aisles and testing of the resulting layouts shows that travel distance is reduced.
Toward an Engineering Discipline of Warehouse Design
Leon McGinnis and Timothy Sprock Warehouses today are complex dynamic engineered systems, incorporating automation, mechanization, equipment, fixtures, computers, networks, products and people, and they can support the flow of tens or hundreds of thousands of different items to enable fulfilling thousands or tens of thousands of orders daily. In that sense, they represent a design challenge that is not terribly different from the design of other complex dynamic engineered systems, such as a modern passenger airplane, an automobile, or a unique building. What is different is that the design of these other complex dynamic engineered systems typically follows some engineering design discipline. Here, we argue for the development of a corresponding engineering discipline of warehouse design.
Understanding Worker Blocking and the Design Process
Russell D. Meller and Lisa M. Thomas Designing a distribution center (DC) is a complex process that consists of answering dozens of interrelated questions. From our perspective, the penultimate question is one of which picking methodology is optimal. To answer this question requires a full evaluation of picking methodologies in terms of their ability to meet the expected throughput requirements of the system. In manual picking systems, as throughput requirements increase, worker interaction and blocking increase as well. This implies that being able to estimate the amount of worker blocking in a system is critical to a design effort. Most of the prior research in this area addresses this point from an analysis perspective. In this paper we provide a design perspective based on the results from a structured simulation modeling study we conducted. We present our main result — a design rule ratio — as well as how this design rule ratio is used within a design process.
Omnichannel Business-to-Consumer Logistics and Supply Chains: Towards Hyperconnected Networks and Facilities
Benoit Montreuil This paper deals with omnichannel business-to-consumer logistics and supply chains. Its key contribution is the conceptualization of hyperconnected network and facility design options for enabling to meet the challenges toward achieving omnichannel logistics efficiently and sustainably while meeting the timely expectations of clients. These design options exploit key Physical Internet concepts. They encompass the transportation, pickup and delivery of ordered goods, the deployment of products across territories to enable fast response to orders, as well as on-demand production of products across networks of facilities while engaging multiple parties. The options range from current practice to prospective ones that are associated to a more mature level of implementation of the Physical Internet. The paper identifies key relative advantages and disadvantages of alternative options, synthesizes strategic insights for industry, and provides research challenges and opportunities.
Outsourced Storage and Fulfillment Facilities to Enhance the Service Capabilities of Shopping Mall Tennants
Zachary Montreuil and Mike Ogle This paper proposes development of an Inventory and Fulfillment Center (IFC) that will provide a new range of outsourced, centralized storage and services for shopping mall tenants. The paper describes the competitive issues faced by shopping mall tenants, the requirements and features of a centralized, outsourced storage and fulfillment facility, and development of an agent-based model in AnyLogic® to demonstrate the concept. The range of research and business challenges introduced by this new concept are also presented.
Logistics Models to Support Order-Fulfillment from the Sea
Jennifer A. Pazour and Ian Shin Sea based logistics use maritime platforms to transfer cargo stored on vessels and delivers them ashore. This chapter describes the motivations and logistical requirements of seabasing. The sea base’s organizational structure, its material handling environment, and the internal cargo flow processes of the TAKE vessel are described. Three seabasing distribution network scenarios — Iron Mountain, Skin-to-Skin Replenishment, and Tailored Resupply Packages — are described and mapped to warehousing and distribution networks, characteristics, and decision problems. Finally, related literature is reviewed and open areas for logistics research to support order fulfillment from the sea are identified.
Integrated Versus Sequential Scheduling and Assignment at a Unit Load Cross-Dock
Arpan Rijal, Marco Bijvank and René de Koster Within a cross-dock, the assignment of trucks to dock-doors and the scheduling of trucks to be processed are two major operational decisions. Conventionally, assignment and scheduling decisions are made sequentially. However, solving the two problems sequentially can lead to sub-optimal solutions because the objectives of the two problems are in conflict with each other. To gain further insights, we create an integrated model which is capable of simultaneously scheduling and assigning trucks at crossdocks. We contrast the integrated model with a sequential model which first schedules trucks for processing and then assigns them to dock-doors. Experiments demonstrate that the integrated model can produce superior solutions, despite that it is computationally more expensive.
Evaluating Transaction Pairing Strategies for Vehicle-Based High-Density Storage Systems
Debjit Roy and Harit Joshi Vehicle-based storage and retrieval systems present an attractive choice for distribution center automation because it provides the flexibility in managing demand fluctuations without affecting transaction throughput times. In this research, we contend that while dual-command cycles can reduce the vehicle travel times for processing transactions, it may not be the best policy for reducing transaction throughput times when transactions arrive at random time instants. We develop stochastic models to test the transaction throughput time performance with multiple pairing strategies and present operational insights.
New Design Guidelines for In-Plant Milk-Run Systems
Thorsten Schmidt, Ingolf Meinhardt and Frank Schulze Tugger trains became a popular means of supply in material handling intensive production systems. In contrast to forklift trucks they interlink a supermarket with multiple delivery locations along a transport route in a milk-run. But the efficiency gain (higher transport capacity, reduced labor costs) has its price: Compared to forklift trucks, planning and dimensioning of in-plant milk-run systems is more complicated. The paper discusses features and drawbacks of a recent standardization approach of the Association of German Engineers (VDI) and highlights the variety of technical restrictions which have to be considered when a milkrun system is designed. It shows, that algorithms can support the design and dimensioning process. It is, however, not feasible to formulate the design task as an ordinary optimization problem which can be handled by a solver without any further interaction.
Maritime Location Decisions for LNG Bunkering Facilities
Reinier C. Schneider and Iris F.A. Vis Liquefied natural gas (LNG) is one of the upcoming fuels to be used for more sustainable shipping activities in the maritime sector. For a widespread adoption by end-users, a refuelling network requiring capital intensive investments, needs to be in place. From a macro perspective of suppliers of LNG, it makes sense to develop the infrastructure at strategic locations that capture as many vessels and ships as possible. The goal of this research is to develop a facility location model that can contribute to the location selection of LNG facilities in a new fuelling network. The new model will fit the maritime sector and specifically the LNG transition. Experiments, with data obtained from expert interviews, have been performed to study facility location decisions in the North Sea areas.
Methodology for Choosing Appropriate Inland Vessels as a Floating Distribution Centers
Andrea Seidlova and David Sourek The primary goal of this paper is to propose a methodology for the choice of an appropriate vessel type under the specific condition of floating distribution center. The right choice of the appropriate type of vessel has a fundamental influence on the long-term success of city logistic projects with water transport. In this paper, we propose a methodology for vessel choice according to its inner layout and technology of handling. The possibility of an application in practice will be defined for each option.
A Heuristic Approach to Storage System Design with Simultaneous Assignment of Goods
Tobias Staab and Willibald A. Günthner In the paper at hand, we analyze the planning of manually operated storage systems consisting of different storage areas. When solving design problems like dimensioning or equipment selection, the assignment of goods to storage areas has to be considered. Based on a general model to describe the geometry and performance calculation of a warehouse, we present a heuristic approach to assign goods to storage areas while simultaneously determining the design of the storage areas in a way that generates the lowest running costs. Our approach aims to support planners during the rough planning stage by providing a quick overview of different solution alternatives. To evaluate our approach, we apply it to an example from industrial practice.
The Storage Replenishment Problem in Rectangular Warehouses
Haldun Süral, Claudia Archetti and Melih Çelik In warehouses, storage replenishment operations involve the transportation of items to capacitated item slots in forward storage area from reserve storage. These items are later picked from these slots as their demand arises. While order picking constitutes the majority of warehouse operating costs, replenishment operations might be as costly in warehouses where pick lists generally consist of only a few lines (e.g., order fulfillment warehouses). In this study, we consider the storage replenishment problem in a parallel-aisle warehouse, where replenishment and order picking operations are carried out in successive waves with time limits. The aim is to determine the item slots that will be replenished and the route of the replenishment worker in each replenishment wave, so as to minimize the total labor and travel costs, and ensure the availability of items at the start of the wave they will be picked. The problem is analogous to the inventory routing problem due to the inherent trade-off between labor and travel costs. We present complexity results on different variants of the problem and show that the problem is NP-hard in general. Consequently, we use a heuristic approach inspired by those from the inventory routing literature. We use randomly generated warehouse instances to analyze the effect of different storage policies (random and turnover-based) and demand patterns (highly skewed or uniform) on replenishment performance, and to compare the proposed replenishment approach to those in practice.
A Lagrangian Relaxation for Capacitated Single Allocation P-Hub Media Problem with Medium Capacity Levels
Ching-Jung Ting and Kuo-Rui Lu In this paper we consider a capacitated single allocation p-hub median problem with multiple capacity levels (CSApHMPMC) in which the decisions are to determine the location of p hubs and their capacity levels, the single allocation of non-hub nodes to hubs in the logistics network. This problem is formulated as an integer programming model with the objective of minimizing the sum of total transportation cost and fixed cost of the selected p hubs with established capacity levels. A Lagrangian relaxation (LR) approach is proposed to solve the CSApHMPMC. The Lagrangian function that we formulated decomposed the original problem into smaller subproblems that can be solved easier. We only solve the CSApHMPMC using Gurobi optimizer for the small sized problems. The experimental results show that the proposed LR heuristic can be an effective solution method for the capacitated p hub median location problem with multiple capacity levels.
Monte Carlo Algorithm to Study Performance Parameters of Shuttle Systems
Wolfgang Trummer and Dirk Jodin Shuttle systems provide alternative solutions in storage technology. Material flow analyses of these systems are extensive due to complex interaction of several shuttles and lift elements. Besides classic cycle time calculation, it also requires advanced analyzing methods. At the Institute of Logistics Engineering (ITL), Graz University of Technology, an innovative, software-based approach has been developed to investigate the performance of shuttle systems. This approach is based on the Monte Carlo method. The software tool takes into account a variety of different systems and operating parameters to reflect operative behavior of shuttle systems.
Analyzing Order Throughput Times in a Milkrun Picking System
J.P. van der Gaast, M.B.M. de Koster and I.J.B.F. Adan E-commerce fulfillment competition evolves around cheap, speedy, and time-definite delivery. Milkrun order picking systems have proven to be very successful in providing handling speed for a large, but highly variable, number of orders. In this system, an order picker picks orders that arrive in real time during the picking process; by dynamically changing the stops on the picker’s current picking route. The advantage of milkrun picking is that it reduces order picking set-up time and worker travel time compared to conventional batch picking systems. This paper is the first in studying order throughput times of multiline orders in a milkrun picking system. We model this system as a cyclic polling system with simultaneous batch arrivals and determine the mean order throughput time. These results allow us to study the effect of different product allocations. For a real world application we show that milkrun order picking reduces the order throughput time significantly compared to conventional batch picking.
Intermodal Hubs in the Physical Internet
Nadeepa Wickramage and Bill Ferrell Distribution of finished goods is currently an effective but inefficient process that consumes significant quantities of fossil fuel to move empty assets. This results in increased costs that are passed to the consumer and unnecessarily increased carbon emissions. The Physical Internet (PI) is focused on shared logistics that could prove to be an important element of next-generation logistics systems. The idea is to store and transport goods in anonymous standard-sized containers so that transportation and warehousing can be efficiently shared by many companies including competitors. If implemented, this idea has the potential of dramatically increasing efficiency thereby reducing fuel consumption and decreasing costs as well as emissions. This paper focuses on one aspect of the PI, intermodal hubs. More importantly, the key difference between the PI hub as imagined in this research and a transhipment facility or breakbulk terminal is that the control is decentralized; hence, this research explores decentralized control of a PI hub through experiments using scenarios and heuristics in an effort to gain some understanding how design and operations impact performance.

2014 IMHRC held in Cincinnati, Ohio
Corporate Hosts: Intelligrated, Inc.

2014 IMHRC Table of Contents and Abstracts

Comparing Industry And Academic Perspectives On Cross-Docking Operations
Paul Buijs and Iris F.A. Vis This paper performs a comparative analysis on the industry and academic perspectives on cross-docking operations. Detailed descriptions are provided for three typical cross-dock settings by means of case illustrations. The purpose of these descriptions is to inspire break-through innovations in future cross-docking research by identifying constraints, decision problems, and performance indicators that are thoroughly anchored in current practice.
Real-Time Dock Door Monitoring System Using A Kinect Sensor
Héctor J. Carlo, Yeiram Martínez and Cristina Pomales-García This study presents a proof of concept where a single Microsoft Kinect sensor is used for automated monitoring of a dock door in real-time. The proposed system will automatically and in real-time: (1) detect when an object breaches the dock door perimeter, and its corresponding speed and direction, (2) count the number of pallets loaded/unloaded to/from a trailer, (3) record the loading/unloading time of each load, and (4) reconstruct an image of every loading/unloading trip at a dock door, irrespectively of the material handling travel speed and direction. Particular emphasis is given to discussing how to extend the proposed concept by using multiple Kinect sensors, the technological challenges for implementation, and the expected benefits of a real-time dock door monitoring system.
Experimentally Investigating The Performance Of Various Order Picking Methods In Different Behavioral Contexts
Jelle De Vries, René De Koster and Daan Stam Three manual picker-to-parts order picking methods (parallel picking, zone picking, and dynamic zone picking) are employed in an experimental warehouse setup and compared in terms of productivity, quality, and job satisfaction. Participants worked in teams and were subject to either an individual-based, or a team-based incentive scheme. Furthermore, the influence of individual participants’ dominant regulatory focus (promotion or prevention) was taken into account. The outcomes show that in parallel picking an incentive system focused on individual performance is beneficial for productivity and quality compared to an incentive system focused on team performance, whereas team-based incentives are more productive in zone picking. These results were more explicitly present for participants with a dominant promotion focus. Participants with a dominant prevention focus picked more productively with team-based incentives in all picking methods. In addition to this, team-based incentives led to a relatively high quality in zone-picking, but a relatively low quality in dynamic zone picking. Our study shows that assigning the right people to the right picking task with a fitting incentive system can substantially cut wage costs without simultaneously harming productivity, quality, or job satisfaction.
An Empirical Study About The Effectiveness Of Lean Empowerment In Warehouses
Kai Furmans and Payam Dehdari Lean Management is well established in production environments. Some empirical evidences are available which suggest that in production systems lean management achieves positive results. For warehousing, some works have already been done, which deal with the application and adaption of lean tools for usage in warehousing. In order to answer the question, whether the application of lean tools leads to a better performance however, no study is available today. Therefore, an empirical study has been conducted, where the effectiveness of lean empowerment has been tested and compared to the performance of warehouse, who continued to work as before.
A Framework For The Robust Design Of Unit Load Storage Systems
Marc Goetschalckx, Pratik Mital and Edward Huang The unit load storage assignment problem determines the assignment of a set of unit loads with known arrival and departure times to a set of unit storage locations in a warehouse. The material handling device(s) can carry at most one unit load at the time. In this research it is assumed that each of the storage locations can be accessed directly without load relocations or rearrangements and that the travel times between the storage locations and from and to the warehousing docks can be computed in advance. The objective is to minimize the total travel time of the material handling device for performing a number of storage and retrieval operations. This type of storage system is in widespread use and implemented in both mechanized and automated systems. It is by far one of the most common storage system architectures for unit loads.

The formulation of this problem belongs to the class of Assignment Problems (AP) but finding the optimal solution for the most general variant is provably hard for large problem instances. A classification of the different variants of the APs for unit loads will be presented. The size of the instance problem is proportional to the product of the number of loads and the number of locations and the number of periods in the planning horizon and is typically very large for real world problem instances. Efficient solutions algorithms only exist for product-based storage policies or for the very special case of a perfectly balanced warehouse for load-based storage policies. However, for load-based storage policies the integrality property is not satisfied in general. This results in very large binary programming problems that to date cannot be solved to optimality. However, the formulations have special structure that can be exploited to design efficient solution algorithms. Properties and the special structure of the formulation will be presented. A specialized compound solution algorithm combines primal and dual approaches and heuristics to reduce the optimality gap. Initial computational experience will be shared. It is anticipated that the solution algorithm can either be directly implemented in commercial warehouse management systems or that it becomes a tool to evaluate the performance of commercially implemented storage policies.

The above formulation is the sub problem in a decomposition algorithm for the design of unit load storage systems that identifies the tradeoffs between efficiency and risk of the performance of the storage system. Different risk measures such as the standard deviation and the downside risk can be used. An example based on realistic data values shows that in this case operator-controlled systems are less expensive and more risky than automated systems. However, if the same level of risk is mandated then the automated system is less expensive.
Clickstream Big Data and “Delivery Before Order Making” Mode For Online Retailers
Yemig (Yale) Gong, Houxuan Xu and Jinlong Zhang Our research is inspired by a leading online retailer using clickstream big data to estimate customer demand and then ship items to customers or hubs near customers by a mode of “delivery before order making” (DBOM) mode. Using clickstream data to obtain advance demand information in order quantities, we integrate the forecasting with a single item uncapacitated dynamic lot sizing problem in a rolling-horizon environment. Using the simulated clickstream data, we evaluate the performance of DBOM mode.
Setting Cutoff Times for Picking Systems With Capacity Degradation
Kevin R. Gue In the new landscape of e-commerce distribution, firms must offer increasingly aggressive delivery promises and then make good on them. These promises often take the form of a cutoff time, such that orders placed before the cutoff time receive premium service (next-day, same-day, etc.) and those placed afterward do not. Later cutoff times are stronger, of course, but the fulfillment system might not be able to process the order before the deadline. How late is too late? We develop a deterministic model to answer this and related questions when the order fulfillment system batches orders for efficient picking operations and therefore exhibits a phenomenon we call capacity degradation.
An Analysis Of Single-Command Operations In A Mobile Rack (AS/RS) Served By A Single Order Picker
Amine Hakim Guezzen and Zaki Sari A Mobile rack Automated Storage and Retrieval Systems (MAS/RS) are picker-to-stock retrieval model which are a variation of the multi aisles AS/RS. This mobile storage system is composed of racks moving laterally on rails so that one can open an aisle between any two adjacent racks, the input/output system, the storage and retrieval (S/R) machine and the computer management system or the control system.

Evaluating an AS/RS could be done using several performance indicators, the two most important ones are: The utilization rate of the S/R machine and the average time necessary to serve storage or retrieval requests (the travel time).

The S/R machine could operate either in single command or in dual command. In a single command, the S/R machine executes either a storage or retrieval operation by cycle. The time necessary to execute a single command is said single cycle time. While in a dual command, the S/R machine executes a storage operation followed by a retrieval operation in the same cycle. The time needed to execute a dual command is said dual cycle time.

In this paper our interest is concerned with the mathematical modeling of single-command operations in a Mobile rack (AS/RS) system. We developed a closed form analytical expression allowing an approximate calculation of the travel time of Mobile Racks-AS/RS.

This expression was compared with an exact discrete expression developed earlier by one of the authors. The models developed in this work are used by Kouloughli et al to determine optimal dimensions of the mobile rack AS/RS that minimize expected travel times.
Order Batching With Time Constraints In A Parallel-Aisle Warehouse: A Multiple-Policy Approach
Soondo Hong, Andrew L. Johnson and Brett A. Peters A commitment of delivery time is critical in some online businesses (De Koster, 2003). An important challenge to meeting customers’ needs is timely order picking which is also relevant to worker safety, item freshness, overall operational synchronization, and reduced overtime. We analyze an order batch picking situation where a trip is constrained by vehicle capacity and must be completed within a specified time. We develop a model which partitions orders to batches to minimize the total travel time such that each trip meets the orders’ time constraints and capacity limit, and also determines a suitable operational policy for each batch. Each policy is characterized by routing method, travel speed, capacity, and pick time. The proposed batching model can simultaneously group orders and can select a best policy among possible policy choices for each batch. To solve the proposed batching procedure, an exact algorithm is implemented based on a branch-and-price method. Our multiple-policy approach experiences 2.1~7.0% reductions in retrieval time compared to a best single-policy approach. The experimental results emphasize that when time constraints are enforced in order batching, a multiple-policy is preferable to a single-policy approach, because allows additional flexibility.
Throughput Analysis Of S/R Shuttle Systems
Georg Kartnig and Jörg Oser Shuttle systems are used in high performance automated storage/retrieval unit load systems. Each storage level is serviced by one transfer car travelling in dual command operation. One buffer slot is located at both ends of each level. This decouples horizontal travel from vertical input/output moves, which in this case requires two independent vertical reciprocating lifts at each end of the aisle. Other systems work with only one lift used in lower throughput applications. The content of this paper is treated in the following sections starting with a problem definition and a literature survey. This is followed by a detailed functional description of the system investigated here. A predictive model with analytical equations is derived for simplified calculations and a comparison with simulation results. A summary, conclusions and an outlook finalize the paper.
Grid Facilities Design: Dynamic Modular Deployment Of Production, Handling And Storage Resources
Suzanne Marcotte and Benoit Montreuil To survive and thrive in a fast-moving environment, facilities must be designed to show adaptability, flexibility and robustness. As some facilities are depicted by heavy and sophisticated equipment costly and hard to displace, others are composed of moveable workstations with highly flexible workers. In most cases, the trade-off is between the cost of redeploying the resources and the excessive cost of material handling and storage incurred by an inefficient deployment of the resources. We propose a design strategy based on (1) conceiving and designing the facility as a stable grid of modules, (2) dynamically deploying production, storage and handling resources to these modules, and (3) dynamically assigning process-product combinations to the modules so as to meet stochastic and dynamically evolving product demand on a rolling planning horizon. We illustrate the strategy as applied to a computer refurbishing and recycling facility.
Integrating Analysis Into A Warehouse Design Workflow
Leon F. McGinnis and Timothy Sprock Supply chain analyses, including those related to material handling systems, are typically purpose-built to answer specific questions and therefore have many different implementations depending on the question, the instance data, and the solver. The purpose-built nature of these models makes it difficult to integrate them into an iterative design workflow. Despite the myriad analysis implementations, the fundamental structure of these systems and their problem domain remains unchanged, suggesting that perhaps analyses could be automatically generated on demand, given an appropriate specification of the particular system to be analyzed. We apply model-based systems engineering (MBSE) methodologies to explore this possibility in the context of functional warehouse design.
Variation In Lifting Behavior During A Highly Repetitive Manual Material Task
Jay P. Mehta and Steven A. Lavender Epidemiological studies have shown an association between manual material handling tasks and low back pain (LBP) (Macfarlane et al. 1997; Hughes et al. 1997; Vandergrift et al. 2012; Lavender et al. 2012). More specifically, manual handling tasks that involve repetitive bending, twisting, carrying or lifting movements have been associated with LBP (Marras et al. 1993; Hoogendoorn et al. 2002; Lotters et al. 2003; Heneweer et al. 2011; Mikkonen et al. 2012; Lavender et al. 2012). In addition, repetitive lifting during manual handling tasks has been associated with muscle fatigue (Dempsey 1998). However, the biomechanical mechanism linking muscle fatigue and back injury development has not been fully investigated. One theory is that muscle fatigue brings about altered behavioral strategies that changes an individual’s exposure to biomechanical risk factors (National Academy Press 2001). Another theory is that momentary muscle substitution patterns result in more variable and less coordinated movements, while still maintaining the same overall behavioral strategy (National Academy Press 2001). With either of these theoretical views, there should be increased variability in biomechanical measures typically used to characterize lifting behavior Larger movement variability may also impose greater loads on the underlying structure.

The aim of the current research was to quantify the biomechanical variation experienced during repetitive asymmetric lifting as often observed in occupational lifting tasks (Marras et al. 1993). Specifically, over the course of a 60-minute repetitive asymmetric lifting task, the behavioral response measures (three-dimensional postural deviations, movement speeds, and spine moments) were hypothesized to exhibit larger peak values, suggestive of a mechanism for injury, as time progressed.
An Experimental Study Of The Impact Of Warehouse Parameters On The Design Of A Case-Picking Warehouse
Russell D. Meller and Lisa M. Thomas The best design for a warehouse is based on its ability to meet the demands placed on the warehouse, which are typically characterized by warehouse parameters like the order profile, inventory requirements, etc. Consequently, these parameters should be considered in the design process. In this paper we characterize the design of a case-picking warehouse with five design variables and identify the warehouse parameters that have the greatest impact in setting the values of these variables. With our analysis, the search for the optimal design can be reduced by limiting the design space considered.
Positioning Automated Guided Vehicles In A General Guide-Path Layout
Abraham Mendoza, José A. Ventura and Subramanian Pazhani The locations of dwell points for idle vehicles in an automated guided vehicle (AGV) system determine the response times for pick-up requests and thus affect the performance of automated manufacturing systems. In this study, we address the problem of optimally locating dwell points for multiple AGVs in general guide-path layouts with the objective of minimizing the maximum response time in the system. We propose a mixed integer linear programming (MILP) formulation for the problem. We also develop a generic genetic algorithm (GA) to find near optimal solutions. The MILP model and GA procedure are illustrated using a two-dimensional grid layout problem. Our computational study shows that the proposed GA procedure can yield near optimal solutions for these test problems in reasonable time.
Modular Design Of Physical Internet Transport, Handling And Packaging Containers
Benoit Montreuil, Eric Ballot and William Tremblay This paper proposes a three‐tier characterization of Physical Internet containers into transport, handling and packaging containers. It first provides an overview of goods encapsulation in the Physical Internet and of the generic characteristics of Physical Internet containers. Then it proceeds with an analysis of the current goods encapsulation practices. This leads to the introduction of the three tiers, with explicit description and analysis of containers of each tier. The paper provides a synthesis of the proposed transformation of goods encapsulation and highlights key research and innovation opportunities and challenges for both industry and academia.
A New Approach For Generating Facility Layouts Using An Algorithmic Approach
Christian Mosblech and Volker Sadowsky In this paper a new approach is described to automatically create layouts for material flow systems. The current research in progress aiming at adopting the methods and algorithms of the Electronic Design Automation to be used in logistics planning is presented. These methods are already applied to create microchips being multiple times more complex than material flow systems while following the same goal: Functional units have to be placed on a predefined area and are linked by connections weighted differently. This basic requirement can be applied to microchip designs as well as material flow systems. The common condition is to create the setup with the smallest connection length possible.

The results are compared to a currently applied computerized method to calculate facility layouts. The overall result of the introduced method is nearly equal to the traditional reference method to create a computerized material flow layout. However, while the new algorithm does all calculations automatically, the traditional method requires manual finishing to achieve a comparable result.

This article thereby shows the potential of the research in progress toward the goal to support logistics planning with a new generation of automated software tools.
Analysis Of Class-Based Storage Strategies For The Mobile Shelf-Based Order Pick System
Shobhit Nigam, Debjit Roy, René De Koster and Ivo Adan Mobile Shelf-based Order Pick (MSOP) systems are gaining significant interest for e-commerce fulfillment due to their rapid deployment capability and dynamic organization of storage pods based on item demand profiles. In this research, we model the MSOP system with class-based storage strategies and alternate pod storage policies using multi-class closed queuing networks. We observe that though closest-open location pod storage policy do not allow to efficiently use the storage spaces in comparison to random location pod storage policy in an aisle, it increases the system throughput for all item classes.
Designing Retail Facilities To Improve Product Exposure
Pratik J. Parikh and Corinne Mowrey A retail facility should effectively engage consumers during their shopping trips if they want to convert demand into purchases. Unfortunately, the complexity of the retailing environment and lack of scientific tools often results in gut-feel approaches experimented in practice. A key factor to retail design often alluded to, but rarely analyzed, is product exposure to the travelling shopper. We define the extent of the shopper's field of vision in order to determine the actual exposure of products experienced by a traveling shopper. In so doing, we can explore the effect rack orientation has on product exposure. Our main contributions include defining product exposure and an approach to estimate it at any point along the travel path. Our results indicate that certain rack orientations result in product exposures as high as 2.5 times that of the traditional 90° orientation.
A Framework And Analysis To Inform The Selection Of Piece-Level Order-Fulfillment Technologies
Jennifer A. Pazour and Russell D. Meller The piece-level order-fulfillment technology selection problem is an important strategic problem that significantly impacts distribution center costs and operations, and is typically solved based on empirical experiences. Given a demand curve and a suite of available piece-level order-fulfillment technologies, we analyze where in the demand curve different order-fulfillment technologies should be applied. To do so, we develop a framework that jointly determines the best combination of piece-level order-fulfillment technologies and the assignment of SKUs to these technologies, which relaxes the sequential-modeling approach of previous research. We validate our methodology with industry data and show that our model provides technology recommendations and SKU assignments that are consistent with successful implementations. Through a set of numerical experiments and statistical analysis, we identify key factors in implementing manual versus automated order-fulfillment technologies and provide observations into the application of different order-fulfillment technology strategies. Finally, we present conclusions and future research directions.
A Two-Level Stochastic Model To Estimate Vessel Throughput Time
Debjit Roy, Vibhuti Dhingra and René De Koster A good estimate of the vessel sojourn time is essential for better planning and scheduling of container terminal resources, such as berth scheduling, quay crane (QC) assignment and scheduling, and fleet size planning. However, estimating the expected vessel sojourn time is a complex exercise because the time is dependent on several terminal operating parameters such as the size of the vessel, the number of containers to be loaded and unloaded, and the throughput of the QCs. The throughput of the QCs in turn depends on the type and number of transport vehicles, number of stack blocks, the topology of the vehicle travel path, the layout of the terminal, and several event uncertainties. To address the modelling complexity, we propose a two-level stochastic model to estimate the expected vessel sojourn time. The higher level model consists of a continuous-time Markov chain (CTMC) that captures the effect of QC assignment and scheduling on vessel sojourn time. The lower level model is a multi-class closed queuing network (CQN) that models the dynamic interactions among the terminal resources and provides an estimate of the transition rate input parameters to the higher level CTMC model. We estimate the expected vessel sojourn times for several container load and unload profiles and discuss the effect of terminal layout parameters on vessel sojourn times.
Experimental Validation Of Travel Time Models For Shuttle-Based Automated Storage And Retrieval System
Zaki Sari, Latéfa Ghomri, Banu Y. Ekren and Tone Lerher In this paper, we aim to validate travel time models for single and dual command cycle displacements of lifts and shuttles in a shuttle-based automated storage and retrieval system (SBS/RS) by using experimental computer simulation. The models under consideration take into account acceleration and deceleration delays. We use ARENA 12 software for the simulation modeling. By simulation, we emulate the real functioning of the system. Therefore, we assume that the results from the ARENA simulation are equivalent to the onsite experimentation. Simulation results are very close to those obtained by analytical travel time models. This shows the high precision of these models to predict operations of SBS/RS.These models can be used at design or operation phases to calculate throughput of the system, to compare between different topologies of SBS/RS or with other types of AS/RS to help decision makers to choose among different alternatives of automated storage systems.
Model And Algorithm For Supply Chain Analysis In Energy-Wood Industry
Thorsten Schmidt and Hannes Hahne Short rotation coppices are a renewable energy concept based on renewable raw material. This paper proposes a model and calculation method which generates valid supply chains using quantified technological and infrastructural constraints.
The Cube Per Order Index Slotting Strategy, How Bad Can It Be?
Peter C. Schuur A well-known and frequently applied policy to assign stock keeping units (SKUs) to (dedicated) storage locations in a warehouse is the Cube per Order Index (COI) slotting strategy. Basically, COI stores an SKU based on how frequently it is picked per unit of stock space required. Fast movers are located close to the Input-Output points. For single command order picking, COI slotting is well-known to minimize order picking travel time. For multi command this is no longer true. An interesting question is: how bad can it be? In this paper we show that there is no limit to this badness. Worst-case behavior of COI is infinitely bad. We construct a worst-case example that proves the following. Given an arbitrary positive integer p, there is (i) a warehouse configuration (ii) a set of SKUs (iii) a set of orders for these SKUs such that slotting these SKUs in the warehouse according to COI leads to an order picking travel time which is p times larger than the order picking travel time produced by an optimal slotting strategy.
Data Model And Software Tools For Modeling Picking Operations
Jeffrey S. Smith When developing simulation models of picking warehouses, it would be helpful to find a way to generate some random pick lists based on a given warehouse layout/configuration. The academic community appeared to be lacking any usable form of “starter code” to perform this function. The best answer was essentially, “we modeled the pick locations as a graph and the picker/vehicle travel using paths through the graph.” The preferred answer would be if the community could provide the data structures, the file read/write code, and any shortest path code based on the data structures that are applied to an order picking model. Unfortunately, it became clear that the code and corresponding data structure specifications were not readily available in any usable format. Yet, the data models and code should be fairly standard and widely available. The request was not focused on proprietary picker-routing, order batching, or WMS code.

With no standards or even de facto standards available, this research began to develop the specifications, data structures, and code. This has been an arduous task that has very little relationship to the actual modeling and research that the code would support. Further, this exact same task has likely been performed hundreds of times by researchers over the years as a precursor to their research/testing. Yet there are no standard data models and associated software tools to represent warehouse configurations and operations. Why?
Optimization Of SKUs’ Locations In Warehouse
David Sourek and Vaclav Cempirek Many companies, which deal with warehousing, optimize their warehouses. The main business of these companies is to offer to customers the warehousing services. It is possible to note that these companies are specialized in warehousing. On the other hand, we can find companies, which have as the main business the selling of goods to customers. These companies use the warehouses too but in a little bit different way. The arrangement of their warehouses can be very unsuitable and convenient for optimization. Our paper is focused on optimization of stored goods’ locations based on market basket analysis. The solution of this task is known as well but not so many authors deal with it. A common problem for many companies is to find sets of products that are sold together. As the source of these information the history of sales transactions is used. The process of the data preparation is mentioned in this paper. The steps described in this paper are applied on real retail store.
Development Of A Heuristics For A Criteria Based Planning Of Pallet Storage Systems
Detlef Spee, Michael Schmidt and Steffen Schieweck Reproducible and quantitative reasoning as the foundation for high-quality planning processes evolves to be key to achieve high quality of speed for logistical processes. This article strives to fulfill this demand by developing a coherent heuristics for the planning of pallet storage systems.

The heuristics uses quantitative approaches provided by the available literature. In case of nonexistent sources, the missing components are developed and integrated. To be applicable in an industrial environment, the comparison and assessment of the created implementation alternatives is mainly monetary based. The heuristics follows a modular structure to achieve adaptability and extensibility. It explicitly does not intend to replace a human designer but to support him during the creation and assessment of high-quality design alternatives.

A concluding case study evaluates the practical applicability of the heuristics and its created solutions. The review of the heuristics turns out to be positive. As a consequence, extension and improvement tasks are proposed.
Modeling And Optimization Of Radio Frequency Identification Networks For Inventory Management
Atipong Suriya and J. David Porter Stock loss and out-of-stocks are outcomes of poorly designed inventory management systems and can lead to significant revenue losses. Inventory management systems (IMSs) based on radio frequency identification (RFID) have the potential to minimize these losses if they are properly designed and deployed. However, the placement of RFID readers to support IMSs is often done on a trial and error basis which is time consuming and results in less than optimal coverage. A methodology to find the optimal location and number of RFID readers in a warehouse facility to ensure a desired level of radio frequency signal coverage was developed in this research. The results show that the proposed methodology works very well when applied to small rectangular facilities and small inverted-T facilities. However, some limitations exist when the method is applied to large facilities.
Comparison Of Alternative Configurations For Dense Warehousing Systems
G. Don Taylor and Kevin Gue In heavily constrained environments requiring very high density storage, traditional aisle-based warehousing may not provide viable options. One feasible manifestation of high density storage systems is the ‘puzzle-based’ system, in which unit loads are moved through the system via manipulation of empty (escort) locations to retrieve desired items. Another option would be the use of movable concentric rings, with escorts being utilized to enable lateral movements of unit loads between the rings. In this paper, the authors present analytical results to compare retrieval time performance for these two types of high density storage systems for randomly demanded items under various assumptions regarding the initial placement of escorts. The paper concludes that the use of movable concentric rings results in significant improvement to retrieval time performance in comparison to rectangular, puzzle-based systems, and further concludes that additional research is warranted.
Integrate Vehicle Routing And Truck Sequencing In Cross-Docking Operations
Ching-Jung Ting and Ting-Dong Cheng Cross-docking is a an important logistics strategy in which freight is unloaded from inbound vehicles and directly loaded into outbound vehicles, with little or no storage in between. This study considers a crossdocking system which combines the vehicle routing problem with crossdocking (VRPCD) for both inbound and outbound operations and truck sequencing problem at docks. The objective is to minimize the logistics center operation costs and transportation costs. We first formulated the integrated problem with a mixed integer programming model. Since VRPCD and sequencing problems are NP-hard, the integrated problem is also an NP-hard problem. We propose an ant colony optimization (ACO) algorithm to solve the VRPCD and sequencing problem by two independent ant colonies sequentially. The proposed ACO is tested with 15 randomly generated instances. The results show that ACO can obtain the optimal solutions in small size instances. We believe the proposed ACO algorithms can be used for practical use for the cross-docking system.
Modeling Conveyor Merges In Zone Picking Systems
Jelmer Van Der Gaast, René De Koster and Ivo Adan In many order picking and sorting systems conveyors are used to transport products through the system and to merge multiple flows of products into one single flow. In practice, conveyor merges are potential points of congestion, and consequently can lead to a reduced throughput. In this paper, we study merges in a zone picking system. The performance of a zone picking system is, for a large part, determined by the performance of the merge locations. We model the system as a closed queueing network that describes the conveyor, the pick zones, and the merge locations. The resulting model does not have a product-form stationary queue-length distribution. This makes exact analysis practically infeasible. Therefore, we approximate the behavior of the model using the aggregation technique, where the resulting subnetworks are solved using matrix-geometric methods. We show that the approximation model allows us to determine very accurate estimates of the throughput when compared with simulation. Furthermore, our model is in particular well suited to evaluate many design alternatives, in terms of number of zones, zone buffer lengths, and maximum number of totes in the systems. It also can be used to determine the maximum throughput capability of the system and, if needed, modify the system in order to meet target performance levels.
Robust Truckload Relay Network Design Under Demand Uncertainty
Hector A. Vergara and Zahra Mokhtari This research addresses the issue of incorporating demand uncertainty in the strategic design of relay networks for truckload transportation. An existing composite variable mathematical model for the design of hybrid relay networks is extended by developing its robust counterpart. The proposed approach considers uncertainty in the number of truckloads to be dispatched between a pair of nodes in the network which is characterized by a symmetric interval around the expected demand value. A two-step heuristic approach is used to solve the robust model. Several numerical experiments are carried out to study the differences between the solutions obtained with the robust approach and those generated by the existing deterministic model. In particular, we were interested in understating how different levels of uncertainty affect total cost of the system and the configuration of the resulting networks. At the end, numerical results are discussed and directions for future research are presented.

2012 IMHRC Table of Contents and Abstracts

Hybrid Heuristics For Infinite Period Inventory Routing Problem
Ronald J. Askin and Mingun Xia In this paper, we address a one-to-many distribution network inventory routing problem over an infinite planning horizon. Each retailer has an independent, random demand, and the distribution center uses capacitated vehicles for routing delivery. The demand at each retailer is relatively small compared to the vehicle capacity. A novel mathematical model is given to simultaneously decide the optimal routing tours to retailers and routing frequencies of each route. Several heuristics are developed to solve large scale instances of the problem.
Functional Design Of Physical Internet Facilities: A Road-Rail Hub
Eric Ballot, Benoit Montreuil and CollinThivierge As part of the 2010 IMHRC, Montreuil, Meller and Ballot enumerated the type of facilities that would be necessary to operate a Physical Internet (PI, π), which they termed, “π-nodes.” This paper is part of a three-paper series for the 2012 IMHRC where the authors provide functional designs of three PI facilities. This paper covers a PI road-rail hub. The purpose of a PI road-rail node is to enable the transfer of PI containers from their inbound to outbound destinations. Therefore, a road-rail π-hub provides a mechanism to transfer π-containers from a train to another one or a truck or from a truck to a train. The objective of the paper is to provide a design that is feasible to meet the objectives of this type of facility, identify ways to measure the performance of the design, and to identify research models that would assist in the design of such facilities. The functional design is presented in sufficient detail as to provide an engineer a proof of concept.
Safe Control Of Manufacturing Vehicles Research Towards Standard Test Methods
Roger Bostelman, Will Shackleford, Geraldine Cheok and Kamel Saidi The National Institute of Standards and Technology‟s Intelligent Systems Division has been researching several areas leading to safe control of manufacturing vehicles to improve automated guided vehicle (AGV) safety standards. The research areas include:

AGV safety and control based on advanced two-dimensional (2D) sensors that detect moving standard test pieces representing humans; Ability of advanced 3D imaging sensors, when mounted to an AGV or forklift, to detect stationary or moving objects and test pieces on the ground or hanging over the work area; and Manned forklift safety based on advanced 3D imaging sensors that detect visible and non-visible regions for forklift operators. Experiments and results in the above areas are presented in this paper. The experimental results will be used to develop and recommend standard test methods, some of which are proposed in this paper, and to improve the standard stopping distance exception language and operator blind spot language in AGV standards.
Using Buffers And Work-Sharing For Minimizing Makespan Of Small Batches In Assembly Lines Under Learning Effects
Yossi Bukchin and E. Wexler The effect of workers’ learning curve on production rate in manual assembly lines is significant when producing relatively small batches of different products. This research analyzes this effect and suggests applying work-sharing among the workers in such an environment to improve the time to complete the batch, namely, the makespan. Work-sharing refers to a situation where adjacent workers help each other in order to reduce idle times caused by blockage and starvation. The effect of work-sharing and existence of buffers on the makespan is examined and compared to a baseline situation, with no work-sharing and buffers. We present mixed-integer linear-programing (MILP) formulations, which minimize the makespan and provides optimal work allocation. A numerical study is conducted and the results along with some operational insights are presented.
A Comparison Of Priority Rules For Non-Passing Automated Stacking Cranes
Héctor J. Carlo, Azaria Del Valle-Serrano, Fernando L. Martínez-Acevedo, Yaritza M. Santiago-Correa and Iris F.A. Vis A recent trend in container ports is to operate dual non-passing Automated Storage Cranes (ASCs) that collaborate to serve storage and retrieval requests from opposite ends of a storage block. Since the ASCs are unable to pass each other, there is an exchange zone that serves as a temporary storage location so that one crane can start a request and leave it to the other crane to complete it. In this study, twelve priority rules are introduced and evaluated to determine which rule minimizes the total makespan for serving all requests, given the sequence in which each ASC will serve the requests. Preliminary results from 12 randomly generated experiments indicate that the priority rules favoring the crane furthest away from the origin of the next request (LonOri) and the longest individual completion times (LonTot) outperformed all other rules in terms of the average percent difference with the best found solution and in terms of the percent of times the priority rule yield the best found solution. Also, combining priority rules AdvFun and LonRem yields the best makespan in 11 of the 12 (91.67%) problem instances tested. Results of this study transcend container ports as it is applicable to any material handling system composed of non-passing MHE and that has pickup/deposit points at the ends of the system.
An MIP Approach To The U-Line Balancing Problem With Proportional Worker Throughput
Andres L. Carrano, Reyhan Erin and Moises Sudit One of the major challenges faced by manufacturing companies is to remain competitive in dynamic environments, where fluctuations in customer demand and production rates require systems capable of adapting in a practical and economical way. A U-shaped production cell is considered one of the most flexible designs for adapting the workforce level to varying conditions. However, re-balancing efforts are time consuming and often require a new work allocation and line design. In this paper, a two-stage MIP model to determine the best cell design under varying workforce levels is proposed. The model seeks to maintain proportionality between throughput and the number of workers. Computational experiments considering various line configurations (up to 19 stations) and workloads (up to 79 tasks) are performed. The results show the proposed algorithm provides excellent results for all small and medium size problems addressed in this study, as well as for certain configurations of large problems. This approach can be used to generate lookup tables of line designs to help with quick reallocation of worker assignments on the shop floor and with minimal disruption.
Collaborative Freight Transportation To Improve Efficiency And Sustainability
Kimberly P. Ellis, R. Steven Roesch and Russell D. Meller Collaborative distribution offers the potential for substantial improvements in freight transportation. As collaboration increases, more loads are available for sharing among transportation service providers, leading to more fully loaded trailers that travel fewer miles and reduce the cost per load on average. In this study, we develop approaches to analyze improvements in key performance measures as collaboration increases in freight transportation. For the data sets analyzed, improvements include a 34% increase in trailer fullness, a 29% reduction in average costs per load, and a 25% decrease in average miles per load. Based on this analysis, collaboration provides substantial improvements for transportation service providers and opportunities for increased driver retention. Drivers would benefit from a better quality of life, more local routes, and more time home with their families. In addition to the economic and social benefits, the environmental benefits include reducing the miles driven and the resulting CO2 emissions.
Humanitarian Logistics – The First Week
Bill Ferrell and Selina Begum Decisions made on material flow during the first week of a natural disaster are critical for victims. Currently, decision makers appears to be making important choices based on experience and intuition with little or no support from quantitative approaches because they do not exist. This research proposes a paradigm and offers two supporting models that will assist decision makers regarding the routing of materials during the first week of a disaster. It explicitly includes information regarding the victims’ needs and the degree to which routes are available in a quantitative way that allows updating as information improves. The paradigm involves the use of information gap theory adapted to the this situation for deciding on the types of supplies to send and the Canadian traveler problem for making decisions on the routes to take.
Analysis Of Material Handling Systems Based On Discrete Time Design Modules
Kai Furmans, Eda Özden, Judith Stoll, Martin Epp, Thorsten Schmidt, Ingolf Meinhardt and Frank Schulze In this paper, we present a method for the performance evaluation of material handling systems in an early planning stage. We are mapping typical material handling elements onto stochastic analytical models. Our intention is to build a toolbox of material handling modules that allows a fast analysis of typical continuous conveying systems. The tool should enable the identification of bottlenecks and the computation of lead time distributions for intra-logistic systems.
A Yard Crane Scheduling Problem with Practical Constraints
Amir Hossein Gharehgozli, Yugang Yu, René de Koster and Gilbert Laporte The problem considered in this paper is to locate storage containers while a set of container storage and retrieval requests are sequenced. Two automated cranes stack and retrieve containers in a single block of a yard. The cranes cannot pass each other and must be separated by a safety distance. Storage containers are initially located at the seaside and landside input/output (I/O) points of the block. Each must be stacked in a specific location of the block, selected from a set of open locations suitable for stacking the storage container. Retrieval containers are initially located in the block and must be delivered to the I/O points. Due to the importance and acceptable waiting times of different modes of transport, requests have different priorities. The problem is modeled as a multiple asymmetric generalized traveling salesman problem with precedence constraints. The objective is to minimize the makespan. We have developed an adaptive large neighborhood search heuristic to quickly compute near-optimal solutions. The numerical experiments show that the solution method can obtain near-optimal solutions.
Robust Material Handling System Design Based On The Risk Versus Cost Tradeoff
Marc Goetschalckx, Edward Huang and Pratik Mital The design and planning of major material handling systems belongs to the class of systems design problems under uncertainty. The overall structure of the system is decided during the current design stage, while the values of the future conditions and the future planning decisions are not known with certainty. Typically the future uncertainty is modeled through a number of scenarios and each scenario has an individual time-discounted total system cost. The overall performance of the material handling system is characterized by the distribution of these scenario costs. The central tendency of the cost distribution is almost always computed as the expected value of the distribution. Several alternatives can be used for the dispersion of the distribution such as the standard deviation and variance. In this study the standard deviation of the cost distribution is used as the measure of the risk of the system. The goal is to identify all configurations of the material handling system that are Pareto-optimal with respect to the tradeoff between the expected value and the standard deviation of the costs; such Pareto-optimal configurations are also called efficient. The final selection of the material handling system for implementation can then be made based on the Pareto graph and other considerations such as the risk preferences of the system owner.
Robust Design of Public Storage Warehouses
Yeming Gong and René de Koster We apply robust optimization and revenue management in public storage warehouses. We optimize the expected revenue of public storage warehouses against the worst cases with a max-min revenue objective, and the decision variables are mainly the number of storage units for each storage type. With the robust design, we can observe worst-case revenue improvement.
A High-Density, Puzzle-Based Order Picking System
Kevin R. Gue and Onur Uludağ We present a new semi-automated system for carton and piece picking operations in the distribution centers. The system features decentralized control based on a message-passing protocol in which cells communicate with their neighbors to determine action in each step. The result is a dynamically changing pick face that provides high sku density and reduced worker travel. We describe an example in which pickers are 20% to 50% more productive than they would be in an equivalent flow rack system.
Simulation-Based Performance Improvement Of A Defense Logistics Warehouse
Sunderesh S. Heragu, Banu Y. Ekren, Gerald W. Evans and John S. Usher Warehouses play a critical role in supply chains. They serve as a vital link between manufacturers and customers. In this study, we investigate performance of the largest warehouse (Eastern Distribution Center - EDC) of the Defense Logistics Agency (DLA) located in New Cumberland, PA and propose a near optimum design for the receiving area of the warehouse to improve its performance. First, we develop a simulation model of the system. Second, we interface an optimization model with the simulation in order to optimize the number of people working at the induction stations. In the optimization model, our goal is to minimize the average cycle time of a material type. Last, we re-design the existing system based on results from the optimization. The simulation and the optimization models are developed through the use of ARENA 13.9 and OptQuest.
Knowledge-Based Methods For Efficient Material Handling Equipment Development
Dirk Jodin and Christian Landschützer This paper focuses on the Knowledge-based engineering (KBE) method and general Knowledge-based technologies for automated design of material handling products. As in vehicle and aerospace engineering this is a widespread technology, the authors try to introduce the main benefits of KBE for the material handling equipment design. The benefits of KBE, as faster and more accurate (safer) product development, customized products and knowledge of employees captured within a knowledge management system, let KBE hold many promising possibilities within, to deal with nowadays demands on engineering development, driven by cost reductions and time shortcuttings.

The authors introduce the actual stage of KBE and its methodologies within engineering development complemented by some critical remarks. There will be shown three different examples of realized KBE projects for material handling equipment, focusing on different levels of automated system the fully automated design of wire rope drums, driven only by some few input parameters specified in a graphical user interface (GUI). Furthermore the automated basic assembly layout for drive components of an AS/RS is shown. The automated layouting and basic design work for shelves and AS/RS within a complete storage systems by a GUI driven input procedure shows the advantages of KBE, for here e.g. in a very fast basic layout design for early stage cost estimation.

The authors will give an outlook of forthcoming work and present ideas to develop an appropriate methodology or working environment for the use of knowledge technologies in material handling design. They focus on the necessary steps and sources to provide, capture and use engineering knowledge and will introduce ideas for software tools to support the use of captured knowledge in automated material handling design.
Analysis Of Parameters Influencing In-Plant Milk Run Design For Production Supply
Eva Klenk, Stefan Galka and Willibald A. Günthner In-plant milk run systems are a transport concept for in-plant material delivery which is becoming more and more applicable especially in the automotive industry. This is due to the system characteristic of providing materials in small lot sizes and with high frequency. As there is a number of different milk run concepts applied and there are several parameters influencing the efficiency and stability of these systems, this paper aims at presenting an overview of common concepts and their properties together with key figures based on an empirical study. The concepts are further analyzed and evaluated with respect to resulting lead times and stability
Performance Trade-Offs In Layouts For Relief Centers
Ananth Krishnamurthy, Sanket Bhat and Debjit Roy At a disaster affected region, relief centers distribute critical supplies and aid to the affected victims. Unlike traditional distribution centers, relief centers experience significant ‘crowd effects’ due to the sudden influx of victims in a confined space. Using knowledge from studies on pedestrian traffic flow, specialized state dependent queuing models are developed to model the flow of victims along the walkways setup at a relief center. The underlying queuing network model is analyzed to derive expressions for the average times that victims experience before they receive the service at the relief center. The research shows that crowd density effects lead to significant increase in congestion and queuing delays underscoring the importance of developing specialized queuing models that assess the impact of congestion effects on alternative layouts of relief centers.
A Multi-Objective Optimization Approach For Designing Automated Warehouses
Tone Lerher, Matej Borovinšek, Iztok Potrč and Matjaž Šraml A multi objective optimization of automated warehouses is discussed and evaluated in present paper. Since most of researchers in material handling community had performed optimization of decision variables with single objective function only (usually named with minimum travel time, maximum throughput capacity, minimum cost, etc.), the multi objective optimization (travel time – cost – quality) will be presented. For the optimization of decision variables in objective functions, the method with genetic algorithms is used. To find the Pareto optimal solutions, the NSGA II genetic algorithm was used. The main objective of our contribution is to determine the performance of the system according to the multi objective optimization technique. The results of the proposed model could be useful tool for the warehouse designer in the early stage of warehouse design.
Refurbishing And Recycling Facilities Design Methodology
Suzanne Marcotte and Benoit Montreuil To design a facility, expected flows between the resources is one of the most important input. Flows are usually calculated given some statistics of previous periods or from the expected demand and the process required. However, in a refurbishing and recycling facility, flows are very fluctuating and not trivial to predict. The quantity produced by such facility not only depends on the demand but also on the supplies which are returned products under guaranty or discarded products after their end-of-use. The uncertainty and the variability on these supplies are often higher than the one on the demand which makes it even more complex to calculate the expected flows. This article contributes a methodology for designing such recycling and refurbishing facilities that are concurrently efficient and robust. It provides an empirical illustration of the methodology through a computer refurbishing and recycling facility case study.
Impact Of Rotated Aisles On Travel Distance In Manufacturing Facilities
Dale Masel and Samantha Hedges The primary metric to evaluate the quality of facility layout is the total distance traveled to move materials around the facility. This distance is dependent on the location of each department in the facility and also on the aisle design. However, most research has only examined the location of the departments in setting up the facility layout and not the location of the aisles. In addition, research on facility layout typically assumes that the aisles will follow the boundary of the facility and are always parallel to the exterior walls of the facility. This assumption reflects the typical practice in how facilities are actually designed, but does not necessarily produce the optimal layout, for minimizing the total distance traveled.

This paper presents a methodology for designing the layout of a facility when the restriction of keeping aisles parallel to walls is eliminated. The methodology allows the main aisles in the layout to be rotated so that they are not parallel to walls, to provide a more direct route between some departments and reduce overall travel distance.
An Object Oriented And Axiomatic Theory Of Warehouse Design
Leon F. McGinnis Despite the conceptual simplicity of warehousing, the development of integrated computational tools for warehouse design has remained an elusive goal. In recent years, there has been some progress toward this goal, with a growing body of research addressing topics as diverse as the design process itself, decision support for specific design decisions, warehouse representation, integrating warehouse representation and analysis, and conceptual approaches for developing integrated warehouse design tool chains. Until now, however, there has not been a suitable warehouse design theory that would provide an integrating framework for all these disparate efforts.

This paper presents an object-oriented and axiomatic warehouse design theory. Key assumptions about the warehouse to be designed are stated as axioms, with appropriate formalisms. Using the axioms and the associated notation, a formal specification of warehouse requirements can be stated, a formal description for a warehouse design can be given, and methods can be developed for testing the warehouse design against the requirements. Moreover, the axioms provide the foundation for identifying essential warehouse design decisions, and formally stating both the criteria for evaluating those decisions and the constraints limiting those decisions.

The paper provides a conceptual and rigorous bridge between the process-oriented research on warehouse design process or workflow, and the mathematically oriented approach to warehouse design reflected in the vast literature on mathematical models and algorithms for specific warehouse design and/or operating decisions
Functional Design Of Physical Internet Facilities: A Road-Based Transit Center
Russell D. Meller, Benoit Montreuil, Collin Thivierge and Zachary Montreuil In their 2010 IMHRC paper, Montreuil, Meller and Ballot proposed a set of facility types that would be necessary to operate a Physical Internet, which they termed “π-nodes.” This paper is part of a three-paper series for the 2012 IMHRC where the authors provide functional designs of three PI facilities. This paper covers a road-based transit center, or road-based π-transit. The mission of a π-transit node is to enable the transfer of π-carriers from their inbound to outbound destinations. Therefore, a roadbased π-transit provides a mechanism to transfer π-trailers from one truck to another. The objective of the paper is to provide a design that is feasible to meet the mission of this type of facility, identify ways to measure the performance of the design, and to identify research models that would assist in the design of such facilities. The functional design is presented in sufficient detail as to provide an engineer a proof of concept.
Functional Design Of Physical Internet Facilities: A Road-Based Crossdocking Hub
Benoit Montreuil, Russell D. Meller, Collin Thivierge and Zachary Montreuil As part of the 2010 IMHRC, Montreuil, Meller and Ballot proposed a set of facility types that would be necessary to operate a Physical Internet (PI, π), which they termed nodes. This paper is part of a three-paper series for the 2012 IMHRC where the authors provide functional designs of three PI facilities. This paper covers a unimodal road-based crossdocking hub designed specifically to exploit the characteristics of Physical Internet modular containers so as to enable the efficient and sustainable transhipment of each of them from its inbound truck to its outbound truck. The objective of the paper is to provide a design that is feasible to meet the objectives of this type of facility, identify ways to measure the performance of the design, and to identify research models that would assist in the design of such facilities. The functional design is presented in sufficient detail as to provide an engineer a proof of concept.
Cycle Time Models For Aisle Changing As/Rs Considering Acceleration/Deceleration
Jörg Oser and Thomas Drobir In multi-aisle AS/RS systems one or more S/R machines serve several aisles with storage racks on both sides of the aisle. An important feature of multi-aisle configurations are aisle changing devices with transfer cars, curve guided aisle switching mechanisms or closed loop rail tracks with curved sections for a two-aisle layout. These systems offer an economic solution, when the number of storage and retrieval transactions is low and storage volume is high. One or several S/R machines each serving a specified number of aisles can be arranged in various layouts to meet the required storage capacity and throughput.
Modeling The Inventory Requirement And Throughput Performance Of Picking Machine Order-Fulfillment Technology
Jennifer A. Pazour and Russell D. Meller Picking machines, also known as remote-order-picking systems, are an example of a stock-to-picker piece-level order-fulfillment technology that consists of two or more pick stations and a common storage area. An integrated closed-loop conveyor decouples the pick stations from the storage area by transporting the needed totes to and from the storage area and the pick stations. We develop a probabilistic model capable of quantifying the inventory differences between order-fulfillment technologies that pool inventory with technologies that do not pool inventory. To determine the throughput of a picking machine, we develop a methodology that incorporates existing analytical models for the picking machine’s subsystems. We present a case study comparing a picking machine to a carousel-pod system to illustrate how a manager could use our methodology to answer system design questions. Finally, we present conclusions and future research.
Understanding And Modeling Sustainability Issues In Facility Logistics
Brett A. Peters and Astrid Garcia Ramos Environmental issues have been the focus of much discussion and debate. In light of environmental concerns, many companies have focused on creating sustainable products and systems. While there is still much disagreement about what is “sustainable,” there is a lot of on-going activity cover a wide range of areas and topics. However, relatively speaking, there is much less attention and activity within a facility logistics context. In this paper, sustainability issues related to facility logistics are explored. An overview of existing related research across a range of areas is provided. Major issues of concern for facility logistics will be discussed and appropriate decision tradeoffs will be characterized to develop potential research issues related to sustainability with the context of facility logistics and closely related topics.
A Mathematical Model For Driver Balance In Truckload Relay Networks
Sarah Root and Hector A. Vergara Driver retention has been cited as one of the primary motivating factors for the implementation of relay networks for full truckload transportation. The strategic design of such networks considering important operational factors such as limitations on load circuity and equipment balance has been previously studied in the literature, however driver scheduling has not been explicitly considered in routing decisions. We present a prescriptive modeling approach that uses mathematical programming in conjunction with a decomposition-based algorithm to select feasible duties that consider current hours-of-service regulations and assign them to drivers domiciled at relay points in the network to cover truckload demands during a given planning horizon. Computational results are presented for randomly generated problem instances along with areas for future research.
Optimal Design Of Container Terminal Layout
Debjit Roy and René de Koster Due to rapid growth in foreign trade using sea vessels, there is a growing focus in improving the infrastructure and operational efficiencies at the container terminals. Particularly, the operational responsiveness of loading and unloading of containers, affects the vessel idle times and profitability of the shipping liners. In this research, we determine optimal stack layout design, which minimizes the container unload times using Automated Guided Vehicles (AGVs). To analyze alternate stack layout designs, we develop integrated queuing network models that capture the stochastic interactions among the container terminal processes (quayside, vehicle transport, and stackside), and provides realistic estimates of expected container unload throughput times.
Design & Modeling Of A Single Machine Flow Rack AS/RS
Zaki Sari and Nadir Hakim Bessnouci In this paper, we aim to introduce a new variation of the flow rack automated storage and retrieval system (AS/RS) using a single machine for storage and retrieval operations instead of two machines. Also, analytical expressions are derived for expected single and dual cycle times of the storage and retrieval machine. For that, randomized storage assignment, and Tchebychev travel are assumed. Two dwell point positions are investigated and compared to determine the best one. Finally an experimental validation using simulation is conducted to verify the quality of the developed.
Process Analysis For Material Flow Systems
Thorsten Schmidt, David Wustmann and Robert Schmaler This paper describes a generic approach for analysis of internal behavior of logistic systems based on event logs. The approach is demonstrated by an example of event data from the simulation model of an automated material handling system (MHS) in a manufacturing company. The purpose of the analysis is the identification of design and operation problems and their causes, prospectively. As a result, the simulation model developer obtains the condensed and ranked information on events. These events describe the internal system behavior with anomalies pointing at either possible problems or capacity reserves.
An Extended Double Row Layout Problem
Alice E. Smith, Chase C. Murray and Xingquan Zuo The double row layout problem (DRLP) seeks to determine optimal machine locations on either side of an aisle, where the objective has been defined as the minimization of material flow cost among machines while meeting machine clearance constraints. In this paper, we extend existing DRLP formulations in two respects. First, we consider the minimization of layout area besides the usual material flow cost objective. Second, we present a mixed integer linear programming formulation that permits non-zero aisle widths. This new formulation also includes new constraints that eliminate layout “mirroring," thus reducing the solution space significantly and thus solution times. Although small-scale problems may be solved optimally by commercial integer programming solvers, solution times are highly sensitive to the number of machines in a layout. A tabu search heuristic is shown to work well for moderately-sized problems. Numerical examples demonstrating the impact of both flow and area objectives, as well as aisle widths, are included.
How To Choose An Order-Picking System
Detlef Spee, Marita Ellinger and Tim Geißen Order-Picking-Systems are the parts of material flow systems, which have the largest variance of deployment alternatives. Since a long time there have been many attempts to systematize the way of identification of the order-picking-system that the “correct” solution can be found.

The previous methods could not be enforced universally. Thus nearly all the material handling system suppliers have their own approach to find the right system, as it turned out at a market analysis of Fraunhofer IML.

With the presented concept of finding a solution, a structured, fast method for system identification shall be presented. A stepwise approach to the system definition reduces the solution space for the next level of detail and thus the processing effort. This procedure should be defined in that way, that it will become a module or component in the integrated approach of warehouse CAD (or the warehouse design workflow) in future. This principle is currently being developed by the GATech and the Fraunhofer IML.
The Order Picking Problem In Fishbone Aisle Warehouses
Haldun Süral and Melih Çelik A recent trend in the layout design of unit load warehouses is the application of layouts without conventional parallel pick aisles. Two examples for such designs are flying-V and wishbone designs for single and dual command operations. In this study, we consider the same layout types under the case of multiple –item pick lists and show that, for both layout types, the routing problem can be solved in polynomial time. We also propose simple heuristics for this problem inspired by those put forward for parallel-aisle warehouses. Our computational results reveal that under certain cases, fishbone design can perform as high as 30% worse than an equivalent parallel-aisle layout, and a modification of the aisle-by-aisle heuristic produces good results compared to other heuristics.
Dock Assignment And Truck Scheduling Problems At Cross-Docking Terminals
Ching-Jung Ting and Amanda G. Rodríguez López In this paper, we consider the integration of dock assignment and truck scheduling problem at cross-docking terminals. The problem is first formulated as a 0-1 integer programming model. Since both dock assignment and truck scheduling problems are NP-hard, its integration is more difficult to solve. Thus we propose reduced variable neighborhood search (RVNS) algorithms to solve the problem. Computational experiments are carried out on four set of instances. The results show that RVNS is capable of finding good solutions in a much shorter computation time when it is compared with optimization solver Gurobi’s solutions.
Estimating Travel Distances And Optimizing Product Placement For Dedicated Warehouses With Manual Picking
Uday Venkatadri and Sachin Kubasad This paper looks at the problem of estimating travel distances for rectangular warehouse sections with manual picking. This study was motivated by a real-life case in the food and beverage industry where case picking occurred in a rectangular section of the warehouse In particular, we are interested in estimating the distance travelled by an order picker whose picking route begins and ends at a single depot. One of the assumptions in many distance approximation papers is that any location is equally likely to be picked. However, this assumption is unrealistic in the case of dedicated warehouse layout, where products are located strategically in order to minimize total distance.

The frequency of accessing a pick location can be estimated from the order history table of a WMS. This in turn can be translated into the probability of accessing certain locations. Under the simplifying assumption that there is no backtracking in the aisles, we build a probability tree to estimate the distance travelled by the order picker.

From a placement point of view, we present three product assignment (or order slotting) heuristics in this paper, namely the North-North, North-South, and Nearest Neighbour heuristics. Our study shows that there is very little variation between the heuristics in terms of travel distance.
Dynamic Slotting Optimization Based On SKUs Correlations In A Zone-Based Wave-Picking System
LI Yingde and Jeffery S. Smith Many exiting slotting methods ignore the picking correlations between Stock Keeping Units (SKUs). In a previous paper, a mix integer program model for dynamic slotting to minimize the pick-wave makespan among all zones under some load balancing constraints was developed. In this paper, we develop an ant colony optimization with slot-exchange policy (ACO-SE) based on SKU correlation to assign the correlated SKUs to the adjacent slots in the same zone. The ACO-SE deposits pheromones between SKUs, uses local and global pheromone trail updates, and controls pheromone accumulation using the Max-Min rule. The main heuristic information is set to the correlation strength and the pick-times are introduced as the assisted heuristic information. A hybrid search mechanism was adopted to improve to global search efficiency. A slot exchange policy was proposed to re-slot the correlated SKUs based on the picks to ignore the proximity of SKUs and to make the farthest SKU for one carton closer to the initial point as far as possible. The promising computational results show that the ACO-SE has perfect convergence and very good CPU time. The solution quality of ACO-SE is always better than the Cube-per-Order-Index (COI), simulated annealing correlation (SA-C) heuristic; it has considerably faster convergence speed than SA-C. The result shows that in zone-based wave-picking system with return touring policy, the exact proximity of SKUs is not critical and that the correlated SKUs can be allocated to any locations along the path from the initial point to the other SKU’s location; the correlation strength has no obvious impact on the picking efficiency, but and correlation probability has significant impact on the picking efficiency.
Open Location Management In Automated Warehousing Systems
Yugang Yu and René de Koster A warehouse needs to have sufficient open locations to be able to deal with the change of item inventory levels, but due to ongoing storage and retrieval processes, open locations usually spread over storage areas. Unfavorable positions of open locations negatively impact the average load retrieval times. This paper presents a new method to manage these open locations such that the average system travel time for processing a block of storage and retrieval jobs in an automated warehousing system is minimized. We introduce the effective storage area (ESA), a well-defined part of the locations closest to the depot; where only a part of the open locations –the effective open locations-, together with all the products, are stored. We determine the optimal number of effective open locations and the ESA boundary minimizing the average travel time. Using the ESA policy, the travel time of a pair of storage and retrieval jobs can be reduced by more than 10% on average. Its performance depends hardly on the number or the sequence of retrievals. In fact, in case of only one retrieval, applying the policy leads already to beneficial results. Application is also easy; the ESA size can be changed dynamically during storage and retrieval operations.
A Study On Storage Allocation In An Automated Semiconductor Manufacturing Facility
Claude Yugma, Stéphane Dauzère-Péres, Ahmed Ben Chaabane, Lionel Rullière, and Gilles Lamiable This paper deals with the allocation of storage capacity in the automated material handling environment of a semiconductor wafer manufacturing facility. The impact of the allocation of stockers to machines in semiconductor fabrication facilities (fabs) is very little studied, although it significantly impacts the efficiency of the Automated Material Handling System (AMHS). After motivating and describing the problem of allocating unitary stockers to machines, a first local approach is discussed. We then propose a Mixed Integer linear Programming model to solve the global problem with two objectives: Minimizing the total maximum travel distance of vehicles and balancing the utilization of unitary stockers. Based on real data, numerical experiments are performed on some small instances. The analysis shows the importance of some parameters and that the two objectives tend to conflict.
Toward Sustainability, High Density And Short Response Time By Live-Cube Storage Systems
Nima Zaerpour, Yugang Yu and René de Koster This paper studies random storage in a live-cube storage system where loads are stored multi-deep. Although such storage systems are still rare, they are increasingly used, for example in automated car parking systems. Each load is accessible individually and can be moved to a lift on every level of the system in x- and y-directions by a shuttle as long as an open slot is available next to it, comparable to Sam Loyd’s sliding puzzles. A lift moves the loads across different levels in z-direction. We derive the expected travel time of a random load from its storage location to the input/output point. We optimize system dimensions by minimizing the expected travel time.
Column Generation For The Container Relocation Problem
Elisabeth Zehendner and Dominique Feillet Container terminals offer transfer facilities to move containers from vessels to trucks, trains and barges and vice versa. Within the terminal the container yard serves as a temporary buffer where incoming containers are piled up in stacks. Only the topmost container of each stack can be accessed. If another container has to be retrieved, containers stored above it must be relocated first. Containers need to be transported to a ship or to trucks in a predefined sequence as fast as possible. Generally, this sequence does not match the stacking order within the yard. Therefore, a sequence of retrieval and relocation movements has to be determined that retrieves containers from the bay in the prescribed order with a minimum number of relocations. This problem is known as the container relocation problem. We apply an exact and a heuristic column generation approach to this problem. First results are very promising since both approaches provide very tight lower bounds on the minimum number of relocations.
Evaluation Of Environmental Benefits Of CHE Emerging Technologies By Using LCA
Nenad Zrnić and Andrija Vujičić In the era of climate change combat transport industry is recognized as a sector with one of the largest environmental footprints. A part of transport industry is container shipping and handling division, which is currently growing with the fastest rate. This massive growth of container sector is due to containers pouring from Asia, mainly from China.

Thus, container port operations are also experiencing significant increase in port emissions. This fact puts port authorities in position to find a way to reduce environmental impact of port operations and at the same time withhold increase in number containers being handled.

In response to the demanding task of reducing emissions and increasing TEU numbers Cargo Handling Equipment (CHE) industry offers variety of solutions. In this paper, state-of-the-art technology for Rubber Tired Gantry (RTG) cranes is being analyzed in order to find out the most eco-efficient solution. A conventional RTG crane is compared to hybrid Eco-RTG with super-cap energy storage system and electrified E-RTG crane. The last two solutions represent the latest trend in CHE industry.

The methodology used to carry out RTG cranes environmental impact comparison is Life Cycle Assessment (LCA) outlined in ISO 14040, as a tool which offers possibility to address entire product's life cycle in a consistent way. The obtained results of RTG cranes LCA are presented in accordance to ISO 14040 principles with the highlight on CML and TRACI impact assessment methods.

Based on the obtained results, the recommendations on reducing environmental footprint of ports are done by necessary improvements on RTG cranes. Since, the objective of this paper is twofold, use of LCA methodology as a tool in the early stage of design is promoted due to its possibility to offer preliminary information and details of processes and materials.

2010 IMHRC held in Milwaukee, Wisconsin
Academic Host: Center for Supply Chain Management at Marquette University
Corporate Hosts: RedPrairie and HK Systems

2010 IMHRC Table of Contents and Abstracts

Empirically-based Warehouse Design: Can Academics Accept Such an Approach?
James M. Apple, Jr., Russell D. Meller and John A. White, Jr. It is our opinion that existing research is not sufficient to support the design of a warehouse. As a result, facility designers that work in practice are left to face the design process with their own methods. These methods are in stark contrast to the analytical models developed in academia in that they are highly based on empirical observations. Supported by a collection of empirical observations, facility designers who work in practice employ an ad hoc design process. We present a design process based on empirical observations and then formalize it so that it can be taught and used. We illustrate the process with an example and discuss ways in which analytical methods can be used to supplement the process and improve the design. We conclude by describing additional work that is required if the design process is to be realized.
A New MILP Approach for the Facility Layout Design Problem with Rectangular and L/T Shaped Departments
Yossi Bukchin and Michal Tzur In this paper we propose a new approach for the facility layout problem (FLP) and suggest new mixed-integer linear programming (MILP) formulations. The proposed approach considers simultaneously the location of the departments within the facility and the internal arrangement of the machines. Two models are suggested, where the first addresses the rectangular department case and the second allows nonrectangular departments defined by an L/T shape. New regularity constraints are developed to avoid irregular department shapes.
Optimizing the Rearrangement Process in a Dedicated Warehouse
Hector J. Carlo and German E. Giraldo Determining the optimal storage assignment for products in a dedicated warehouse has been addressed extensively in the Facility Logistics literature. However, the process of implementing a particular storage assignment given the current location of products has not received much attention in the existing literature. Typically, warehouses use downtime or overtime to remove products from their current location and move them to the suggested location. This work presents the Rearrange-While-Working (RWW) policy to optimize the process of rearranging a dedicated warehouse. The RWW policy seeks to relocate products in a warehouse from the initial arrangement to the optimal arrangement while serving a list of storages and retrievals. This study considers three scenarios: (1) when there is only one empty location in the warehouse and the material handling equipment (MHE) is idle (i.e. reshuffling policy); (2) when there is only one empty location in the warehouse under the RWW policy; (3) when there are multiple empty locations in the warehouse under the RWW policy. In the first case, the MHE can make any movement desired as it is idle. In the other cases, the movements correspond to a list of storages and retrievals that need to be served. In these cases it is assumed that products can only be moved when they are requested. After being used, they are returned to the warehouse. Several heuristics are presented for each scenario. The proposed heuristics are shown to perform satisfactorily in terms of solution quality and computational time.
Accidents will Happen. Do Safety Systems Improve Warehouse Safety Performance René B.M. de Koster, B.M. Balk, I. Davelaar and M. Martens Safety is becoming more and more an issue in warehouses. In the literature, effective measures leading to increased occupational health and safety have hardly been researched. Most research focuses on the impact of perceived safety-related leadership of managers and worker safety consciousness on ‘safety climate’ and workers’ safe behavior. We have carried out exploratory research into which measures really improve the safety performance of a warehouse. We particularly focus on the effects of (1) safety-related work procedures, (2) safety leadership, and (3) workers’ safety consciousness. Based on a survey we show that safety leadership and safety-related work procedures significantly drive worker safety consciousness, which in turn positively impacts safety performance.
Spiral Facility Layout Generation and Improvement Algorithm Fahrettin Eldemir and Hatice E. Sanli This study is concerned with construction and improvement of a facility layout heuristic called Spiral Facility Layout Generation and Improvement Algorithm (SFLA). The algorithm starts with positioning departments from center point and continues like a hologram from center to outside. The aim of any facility layout algorithm is to better allocate the departments within facility. SFLA is compared with the existing space filling curve methods, MCRAFT and MULTIPLE, that are available in the literature. To form an initial spiral curve, a block system is used, like the bands for MCRAFT. The width and length of the blocks are given by user and departments are formed according to these values and placed around the spiral curve.

The initial layout can be selected either randomly or with a method which is called enhanced initial layout. Enhanced initial layout find the highest related department and put it into center and then add the other departments according to their relationships with the previous one.

20 departments data by Armour and Buffa (1963) have been used to test the performance of the SFLA. For bandwidth 4, SFLA gave better results than both MULTIPLE and MCRAFT. For the same flow data, 200 initial sequences selected randomly. Then initial layouts generated from these random sequences. Using pair wise exchange improvement methods both MCRAFT and SFLA layouts are improved. For bandwidth 4, SFLA often yielded better results than MCRAFT. Significance of the results is statistically tested.
Freight Sequencing to Improve Hub Operations in the Less-Than-Truckload Freight Transportation Kimberly P. Ellis, Xiangshang Tong and Amy Brown Greer In less-than-truckload freight transportation, hub operations affect the service levels that carriers are able to provide their customers. This paper focuses on improving the efficiency of hub operations by reducing freight handling time and cost. Specifically, the freight sequencing problem (FSP) is investigated to determine the freight unloading and loading sequence that minimizes the time for dock workers to transfer shipments from origin trailers to destination trailers. The FSP is modeled as a Rural Postman Problem (RPP) and three algorithms are compared: trailer-at-a-time, nearest neighbor, and balance-and-connect. Using five industrial data sets, the results demonstrate the effectiveness, advantages, and disadvantages of the approaches.
Monitoring Inventory Control Accuracy with Statistical Process Control John R. English, Kyle Huschka, Todd Easton and Andrew Huschka Inventory accuracy is critical in most industrial environments such as distribution, warehousing, and retail. Many companies use a technique called cycle counting and have realized outstanding results in monitoring and improving inventory accuracy. The time and resources to complete cycle counting are sometimes limited or not available. In this work, we promote statistical process control (SPC) to monitor inventory accuracy. Specifically, we model the complex underlying environments with mixture distributions to demonstrate sampling from a mixed but stationary process. For our particular application, we concern ourselves with data that result from inventory adjustments at the stock keeping unit (SKU) level when a given SKU is found to be inaccurate. We provide estimates of both the Type I and Type II errors when a classic C chart is used. In these estimations, we use both analytical as well as simulation results, and the findings demonstrate the environments that might be conducive for SPC approach.
Forklift Routing in Warehouses Using Dual-Commands and Stackable Pallets Bill Ferrell and Ahmed Hassan This research determines time optimal routes for loading and picking pallets that can be stacked on top of each other during transport in a manual warehouse that only contains full pallets and utilizes single deep storage. This research was motivated by the fact that we are seeing this situation on an ever increasing basis, particularly in warehouses that supply parts to automotive assembly. In practice, forklift drivers have developed strategies to take advantage of this opportunity but to our knowledge there is no literature that addresses this problem rigorously. The important features of this work are that a time based mathematical model is required because the time spent stacking and unstacking pallets can be significant and a modeling approach to including stacking had to be developed. The basic models are included here with examples and insights into future work required for applicability to a wider range of users.
Gains in the Life-Cycle of Adaptable, Self-Organized Material Handling Systems Guido Follert, Jan R. Nopper and Michael ten Hompel Compared to conventional material flow controls, self-organized material handling systems and the Internet of Things in facility logistics promise several advantages during the life-cycle. Most important is the increased adaptability in case of expansions or modifications due to a consistent modular design; this also promotes an increased robustness due to clearly defined interfaces and a decreased complexity of each module. The use of RFID technology increases the availability of real-time data about the system and the transported units. However, the introduction of self-organized material handling systems also causes costs, e.g. for necessary RFID tags and readers. Against this background, it is unsatisfactory that the increased adaptability as the main advantage of these systems is hard to grasp. This paper proposes a methodology to analyze the advantages of adaptability in facility logistics during the life-cycle of a material handling system and illustrates its usage. The proposed methodology is based on a dynamic optimization of payoffs during the life-cycle; thereby, all payoffs which are influenced by the adaptability of the material-handling system are included; therefore, the methodology allows to consider the adaptability of all material handling systems.
Plug-and-Work Material Handling Systems Kai Furmans, Frank Schönung and Kevin R. Gue One disadvantage of automated material handling systems is their relative Inflexibility: once racks are installed and conveyors are laid, making even minor changes to a system can be cumbersome and expensive. However, recent progress in the capabilities and cost of basic system components, such as controllers, drives, and sensors, has made possible a new class of material handling systems having a much higher degree of flexibility. We propose underlying design principles for such systems and describe several prototype “plug-and-work" systems, which feature decentralized control and ease of reconfiguration.
Strategic Design of a Robust Supply Chain Marc Goetschalckx and Edward Huang The strategic design of a robust supply chain has as goal the configuration of the supply chain structure so that the performance of the supply chain remains of a consistently high quality for all possible future scenarios. We model this goal with an objective function that trades off the central tendency of the supply chain profit with the dispersion of the profit as measured by the standard deviation for any value of the weights assigned to the two components. However, the standard deviation, used as the dispersion penalty for profit maximization, has a square root expression which makes standard maximization algorithms non applicable. The focus in this article is on the development of the strategic and tactical models. The application of the methodology to an industrial case will be reported. The optimization algorithm and detailed numerical experiments will be described in future research.
A Real-Time Picking and Sorting System in E-commerce Distribution Centers Yeming Gong, Erik M.M. Winands and René B.M. de Koster Order fulfillment is the most expensive and critical operation for companies engaged in e-commerce. E-commerce distribution centers must rapidly organize the picking and sorting processes during and after the transaction has taken place, with the ongoing need to create greater responsiveness to customers. Sorting brings a relatively large setup time, which cannot be well admitted by existing polling models. We build a new stochastic polling model to describe and analyze such systems, and provide approximate explicit expressions for the complete distribution of order line waiting time for polling-based order picking systems and test their accuracy. These expressions lend themselves for operations and design operations, including deciding between “pick-and-sort” or “sort-while-pick” processes, and warehouse performance evaluation.
The Human-Centric Warehouse Kevin R. Gue In theory and in practice, the objective of warehouse design has long been to meet operational requirements of throughput and service performance at minimum cost. A natural product of this “bottom line” approach has been warehouse buildings and operations ill-suited to the humans who spend their working lives interacting with them. We offer an explanation for the current approach, and argue that a new approach— a new paradigm —is needed. We then describe a comprehensive approach to warehouse design that includes human well-being as a primary goal.
Merge Problems with High Speed Sorters Dirk Jodin and Andreas Wolfschluckner To meet customer demands on distribution the necessity of sorting systems for unit loads in regard to quantity and capacity is growing up constantly.

Due to that, manufactures offer powerful sorting machines with a capacity up to 15.000 pieces per hour. Those machines are for instance Crossbelt or Tilt-Tray sorters. The output of these sorters is calculated by the equation of throughput.

The point of interest is the merge, referring to the overall efficiency of the single sorter. The motion sequence during merging has to be absolutely quick and precise to achieve exact positioning using minimized distances of the single items and high velocities on the main conveyer. Actual research activities at the Institute of Logistics Engineering at the Technical University of Graz will provide improved solutions for this problem.

The mathematical and mechanical models are representing a way how to describe and optimize the merging of single piece items. A basic example will show the transactions while merging in detail and the results will be discussed. Furthermore a layout proposal to simplify and increase the merging of goodswill be presented.
Evaluating the Effect of Operational Conditions and Practices on Warehouse Performance Andrew L. Johnson Analyzing warehouse performance across different environments is critical to improving overall productivity and reducing costs. Although two-stage DEA estimators have been shown to be statistically consistent, the finite sample bias of DEA in the first stage carries over to the second-stage regression, which causes bias in the estimated coefficients of the contextual variables. The bias is particularly severe when the contextual variables are correlated with inputs. To address this shortcoming, we apply insights from Johnson and Kuosmanen (2010), who demonstrate that DEA can be formulated as a constrained special case of the Convex Nonparametric Least Squares (CNLS) regression to develop a new semi-parametric one-stage estimator. The new model is applied to a set of warehouses to illustrate its performance.
Dispatching Vehicles Considering Uncertain Handling Times at Port Container Terminals Kap Hwan Kim and Vu Duc Nguyen This paper considers the problem of vehicle dispatching at port container terminals in a dynamic environment. The problem deals with the assignment of delivery orders of containers to vehicles while taking into consideration the uncertainty in the travel times of the vehicles. Thus, a real-time vehicle dispatching algorithm is proposed for adaptation to the dynamic changes in the states of the container terminals. To evaluate the performance of the proposed algorithm, a simulation study was conducted by considering various values of decision parameters under the uncertainty in travel times. Further, the performance of the proposed algorithm was compared with those of heuristic algorithms from previous studies.
Blocking Effects on Performance of Warehouse Systems with Autonomous Vehicles Ananth Krishnamurthy, Debjit Roy, Sunderesh Heragu and Charles Malmborg Distribution centers are under increasing pressure to adopt material handling systems that offer greater flexibility to improve cycle time and throughput capacity in the transfer of unit loads in their high density storage areas. Autonomous Vehicles-based Storage and Retrieval Systems (AVS/RS), have been shown to hold significant promise in this context. In these systems, loads are transferred by autonomous vehicles. Vehicles support horizontal load movement along aisles and cross-aisles within a tier, and lifts support vertical movement between tiers. Existing research in AVS/RS, do not explicitly account for the potential blocking of vehicles and lifts while they are processing transactions. Blocking could occur when multiple vehicles use the same cross-aisle or aisle to process transactions. These blocking delays could significantly impact throughput capacity and cycle times. In this research, protocols are developed to address vehicle blocking and a simulation model is proposed to analyze system performance and quantify the effect of blocking.
Designing Automated Warehouses by Minimising Investment Cost Using Genetic Algorithms Tone Lerher, Iztok Potrč and Matjaž Šraml The successful performance of the automated storage and retrieval systems is dependent upon the appropriate design and optimization process. In the present work a comprehensive model of designing automated storage and retrieval system for the single- and multi-aisle systems is presented. Because of the required conditions that the automated storage and retrieval systems should be technically highly efficient and that it should be designed on reasonable expenses, the objective function represents minimum total cost. The objective function combines elements of layout, time-dependant part, the initial investment and the operational costs. Due to the non-linear, multi-variable and discrete shape of the objective function, the method of genetic algorithms has been used for the optimization process of decision variables. The presented model prove to be very useful and flexible tool for choosing a particular type of the single- or multi-aisle system in designing automated storage and retrieval systems. Computational analysis of the design model indicates the model suitability for addressing industry size problems.
Determining Warehouse Storage Locations Assignments Using Clustering Analysis Dale Masel and Carlos Egas The methodology used to assign products to a storage location in a warehouse can have a significant impact on the amount of time required to retrieve all of the items needed to fill an order. This paper describes a methodology that uses a clustering approach to determine storage assignments, where the metric of the strength of the relationship between two stock-keeping units (SKUs) is the number of times that the SKUs appear in the same order. Clustering is performed to maximize the frequency with which SKUs in the same cluster are ordered together. In testing, the clustering assignments were compared to a demand-based assignment strategy and showed a reduction of 20-30% in the number of aisles visited to retrieve orders.
The Future of Modeling in Material Handling Systems Leon F. McGinnis Today, when we talk about “modeling” in the context of material handling systems, invariably we are referring to a mathematical or computational model for analyzing some aspect of the system, such as its throughput rate, response time, cost of ownership, required storage capacity, etc. Creating these kinds of models requires considerable knowledge in at least two domains—the material handling system domain, and the analysis methodology domain—and considerable skill in the “art of modeling” in order to express the former in the terms of the latter. The results can be somewhat ad hoc—e.g., two different modelers are likely to create two somewhat different simulation models of exactly the same material handling system. In the past, the situation in software development was very similar, with individual programming experts idiosyncratically driving software development. Over the past twenty years, however, computer scientists and software engineers have created a radically different approach to the process of software “modeling” called Model Driven Architecture, or MDA, that is used to create software for standard applications. The thesis of this paper is that MDA can be adapted to the kind of modeling done to support design and operational decision making in material handling systems. The paper describes MDA technologies in the context of material handling system modeling, and explains how adapting this approach to our context will transform the way we do research and the way material handling systems are analyzed and designed in practice.
Optimizing Distribution Center Configuration: A Practical View of a Multi-Objective Problem Russell D. Meller and Lisa M. Thomas The shape of a distribution center, as well as whether dock locations are on one side or two sides of the facility, impacts measures like travel distances and the number of dock locations that may be utilized. Thus, for a required number of pallet locations, there are multiple combinations of distribution center shape and dock configurations that should be evaluated against multiple measures. We have developed a practical model for making such evaluations and illustrate the model with data reflective of a partner in the Center for Engineering Logistics and Distribution.
A Flowshop Scheduling Problem with Transportation Times and Capacity Constraints Abraham Mendoza, José A. Ventura and Kwei-Long Huang Although there are numerous methodologies and research studies on machine scheduling, most of the literature assumes that there is an unlimited number of transporters to deliver jobs from one machine to another for further processing and that transportation times can be neglected. These two assumptions are not applicable if one intends to generate an accurate schedule for the shop floor. In this research, a flowshop scheduling problem with two machines, denoted as M1 and M2, and a single transporter with capacity c is considered. The main focus is on the development of a dynamic programming algorithm to generate a schedule that minimizes the makespan. The transporter takes t1 time units to travel with at least one job from machine M1 to machine M2, and t2 time units to return empty to machine M1. The computational complexity of the proposed algorithm is shown to be O(n^3).
Towards a Physical Internet: The Impact on Logistics Facilities and Material Handling Systems Design and Innovation Benoit Montreuil, Russell D. Meller and Eric Ballot Aiming for a radical sustainability improvement, the Physical Internet has the potential of revolutionizing the fields of material handling, logistics, transportation and facilities design. It exploits the enabling concept of standardized, modular and smart containers as well as the universal interconnectivity of logistics networks and services. Its underlying paradigm shift creates a tremendous breakthrough innovation opportunity for the material handling and facility logistics community in terms of equipment, systems and facility design and operation. This paper provides a primer overview of a key subset of the physical elements serving as the foundation of the Physical Internet infrastructure, classified in three categories: containers, movers and nodes. Each element introduced is characterized and illustrated to enable visualization of their innovative nature. The paper helps uncover a wide variety of potent research avenues.
The Effect of Interaction Between the Production System (PS) and a Looped Conveyor-Based Material Handling System (LCMHS) in a Manufacturing Facility Dima Nazzal and Vernet Lasrado In this paper, we provide empirical evidence that shows the effect of the interaction between the production system (PS) and a looped conveyor-based material handling system (LCMHS) in a manufacturing facility. A rudimentary simulation model captures the interaction between the two systems. Varying several key factors, we test for a statistically significant difference in the work in process (WIP) of the production system with and without the LCMHS to find if the squared coefficient of variation (SCV) of the interarrival time distribution to the PS is affected. The results suggest the need to model the interaction between the two systems in order to obtain a more representative estimate of the WIP in a manufacturing facility.
Drive and Motion Design in Material Handling Equipment Jörg Oser and Christian Landschützer Drives account in many cases up to one third of the costs of material handling equipment. This fact justifies a closer look to important drive and motion issues. Typical design criteria for drives are energy and power consumption, wear, heat and noise generation. Engineering design activities start with the generation of the system configuration, that is to make appropriate topological decisions where to locate the drives in the equipment structure. These decisions define to a great extent the functional quality of the mechanical structure and the distribution of forces in the power train. For early design stages an elasto-kinetic model is developed, which is later enhanced by a more detailed simulation model. Another important issue is the definition of high quality motion profiles defined by selected velocity-time relationships.
Distribution Planning Considering Warehouse Decisions Pratik J. Parikh, Xinhui Zhang and Bhanuteja Sainathuni Modern supply chains heavily depend on warehouses for rapidly fulfilling customer demands through retail, web-based, and catalogue channels. The traditional approach that considers warehouses as cost-centers has affected the profitability of numerous supply chains. A lack of synchronization between procurement and allocation decisions causes warehouses to scramble for resources during peak times and be faced with under-utilized resources during drought times. Warehouses, however, have emerged as service-centers and it is imperative that warehousing decisions be an integral part of supply chain decisions. In this paper we propose a mixed-integer programming model to integrate warehousing decisions with those of inventory and transportation to minimize long-run distribution cost. Preliminary experiments suggest a sizeable reduction in the level and variance in the warehouse workforce requirements. A cost savings ranging between 2-6% is also realized.
Analysis of Picker Blocking in Narrow-Aisle Batch Picking Brett A. Peters, Soondo Hong and Andrew L. Johnson This study analyzes impacts by batch picking on picker blocking in narrow-aisle order picking, and determines appropriate batch formations for a better order picking throughput. We present multiple-pick analytical models to offer insights about picker blocking in batch picking. Several simulation studies over a variety of batching situations scrutinize order batching situations which give throughput benefits in narrow-aisle configurations by satisfying the analytical results. Our results highlight three findings for narrow-aisle batch picking processes: 1) variation of pick probability in batch picking is inevitable and is a primary driver of picker blocking; 2) a near-optimal distance-based batching algorithm can experience less picker blocking than expected, because it reduces both the number of aisles visited and the variation in the number of picks per aisle; and 3) the sorting strategy itself (i.e., pick-then-sort or sort-while-pick) causes varying amounts of congestion.
Design of Cross-Chain Internet Order Fulfillment Centres Kees Jan Roodbergen, Iris. F.A. Vis and Jaap Boter Many consumers have embraced the option of ordering via the Internet, which has resulted in an enormous increase in direct orders compared to the times when direct ordering was done by catalogue and phone. The fulfillment process in the supply chain is an important factor for these consumers impacting how long they must wait between ordering and delivery. This fact has significantly increased the importance of the back-end fulfillment process. We present a novel supply chain design to enable cross-chain coordination of order fulfillment operations for internet sales. Shared warehousing facilities are used more and more to achieve competitive advantage. This situation asks for new models to enable a smooth warehousing process for each web shop, but at the same time to ensure overall efficiency and effectiveness. This paper introduces a layout model for shared operations under one roof by simultaneously optimizing the overall facility layout and the area layout.
Multiple-Grooved Magnetic Traction Sheaves Thorsten Schmidt, Thomas Leonhardt and Martin Anders This paper discusses a new and innovative form of traction sheaves, which can be used in many hoisting applications that are driven with wire ropes in material handling systems. By use of high performance permanent magnets, integrated in the periphery of traction sheaves, a significant increase of driving capacity is achieved. In this special case it concerns high energy magnets consisting of rare earth materials known as NdFeB magnets. These magnets provide a 20 times higher force than conventional magnets. The results of the research show that the driving capacity of traction sheaves with a round groove design can be increased by around 25% through the use of NdFeB magnets. By this means it is possible to combine high driving capacity with low wearout for the wire rope. Conventional traction sheaves with high driving capacity, such as V-groove sheaves go along with fast wire rope deterioration because of the high local pressure stress on the rope.
WEBASRS – A Web-Based Tool for Modeling and Design of Abstract Unit-Load Picking Systems Jeffrey S. Smith and Sabahattin Gokhan Ozden This paper describes a web-based tool that supports the modeling and design of abstract unit-load picking systems. The term “abstract” implies that the model is not specific to any equipment or vendors’ products, but, instead, focuses on the generic system components such as pallets, racks, slots, forklifts, cranes, etc. that comprise typical unit-load picking systems. The objectives of the tool are to support the design of an AS/RS-based or a manual forklift-based picking system based on a set of design parameters and to be able to convert from an AS/RS design to a flat warehouse design and vice versa. The research objective is to design the formal model (the data structure and operational description) that supports the conversion from one type to the other and supports the generation of static and dynamic analysis models and the recording of the analysis results. The web implementation uses a mix of XML, HTML, JavaScript and PHP and implements two existing analysis methodologies from the literature.
Design of Warehouse for Material, Which is Non-Uniform and Difficult to Handle David Sourek and Andrea Seidlova Main goal of this paper is warehouse layout optimization with emphasis on maximum space utilization. It is about rack system design, which leads to capacity maximization of warehouse for non-uniform material with heavy weight. The warehouse system has to enable at least partial automation (the possibility to use AS/RS system), minimization of service time and minimization of man effort. Initial constraining conditions for warehouse design in our case are warehouse area, maximum load of racks or shelves, maximum allowed floor load and existing location of arrival road for material loading and unloading.
Optimization of Costs and Service Level by Choosing the Best Customer Order Decoupling Point Detlef Spee and Michael Schmidt The customer order decoupling point (CODP) as the link in the sup-ply chain between processes based on uncertain information, such as sales forecasts, and certain information in the form of customer orders is crucial for production efficiency, storage costs, and the quality of logistics. This paper focuses on developing an approach for identifying a CODP that provides the highest potential for achieving business objectives. Within the context of the above-mentioned topics, this paper focuses on the inter-relations and the tradeoffs that have to be made when positioning the COPD. The goal of the paper is to present an effective and chronological sequence of tasks, analyses and methods, criteria, and indicators that can help a production planner determine the CODP. In the first part of the pa-per, the factors affecting the positioning of the CODP (internal and external to the company) and their characteristics are identified. Based on extensive literature research, these factors are then mapped to manufacturing concepts, such as make-to-stock, assemble-to-order, and make-to-order. The factors that lead to moving the CODP are also identified and used as the foundation for the development of an iterative procedural method with four stages. In stage 1, the as-is state and current goals are captured and suitable CODPs are identified for the above-mentioned logistic factors. In stage 2, the products are combined into groups to reduce the complexity of the analysis and a procedure for accomplishing this is proposed. The number of suitable CODPs is reduced in stage 3 and the factors that influence the positioning of CODPs are taken into consideration. In stage 4, a financial and qualitative evaluation is performed on the various CODPs.
Investigating Possible Synergies in Intermodal Operations with Truck and Rail G. Don Taylor, Mia K. Burns and Gary L. Whicker As the trucking industry continues to examine ways to provide better service at lower cost, many companies are more heavily utilizing intermodal (IM) strategies between truck and rail, especially for those loads that are relatively non-critical in terms of delivery time requirements and that have longer lengths of haul. As IM business grows, supporting dray infrastructure naturally develops around IM rail yards. What is unknown is whether it is best to have a dedicated set of drivers performing dray operations or if efficiency and cost savings can result when utilizing a joint driving fleet to concurrently support IM and traditionally dispatched truckload freight transportation. This paper describes a set of experiments utilizing a comprehensive discrete-event system simulation model and historical data from J.B. Hunt Transport to determine whether or not operating synergies exist when IM dray operations are integrated with local, regional, and long-haul trucking operations. Performance metrics of interest to drivers, customers, and trucking companies are utilized to ensure that the research addresses issues of importance to all constituencies. The results show that there is a trade-off between different performance variables when combining operations, but that generally speaking synergies do exist when considering the needs of professional drivers. Results are more mixed with respect to the needs of carriers and customers, but the authors reach the conclusion that the positive aspects of combining OTR and IM dispatching activities outweigh the negative. Because the evaluative simulation model itself is considered to be a major contribution, it is also described in some detail herein.
Using Data Development Analysis to Evaluate the Performance of Third Party Distribution Centers Ching-Jung Ting and Hsueh-Lin Fang There has been considerable interest worldwide in last few years in the growth of third party logistics (3PL) providers. 3PL distribution center (DC) enables firms to achieve reduced operating costs, increased revenues, and to focus on their core competence. This research aims to find the key performance indicators through a survey of a set of DCs and then evaluate their efficiency over the period 2005-2007 using data envelopment analysis (DEA) models based on selected performance indicators as inputs and outputs. Three inputs and two outputs for all DCs from the surveyed performance indicators were selected in this study. DEA is a non-parametric linear programming technique used to evaluate the efficiency of decision making units (DMUs) where multiple inputs and outputs are involved. We adopted both the input-oriented CCR model and the BCC model that were designed to derive weights instead of being fixed in advance and handle positive inputs/outputs. A Malmquist productivity index (MPI) analysis further evaluates efficiency change and productivity growth between two time points. Our empirical results show that scale inefficiency is the major reason for the inefficient DMUs. For the future research, more DC data should be collected and different DEA models could be applied for other benchmark studies.
Determination of Operational Parameters for an Efficient Container Service in the Port of Guaymas J. René Villalobos and Cesar Meneses The port of Guaymas is located in Sea of Cortez in the Northern Pacific Coast of Mexico. Its hinterland is basically the Northwestern region of Mexico and the Southwestern United States. The Port currently focuses on bulk and liquid cargo and does not provide container services. In this paper, we explore some of the characteristics that a container service should have to be competitive in servicing the needs of the regional industry. Since the study deals with port selection decision from the industry’s point of view, we introduce a port selection model based on a Total Landed Cost (TLC) metric. The findings show that under the right conditions, the Port of Guaymas is an attractive option for the companies located in its hinterland.
Control Policies for a Dynamic Storage System with Multiple Lifts and Shuttles Iris F.A. Vis, Hector J. Carlo, Bruno van Wijgaarden New types of Automated Storage and Retrieval Systems, able to achieve high throughput levels, are continuously being developed and require new control polices to take full advantage of the developed system. In this paper we study a dynamic storage system as developed by Vanderlande Industries consisting of a conveyor, two lifts, multiple transfer shuttles, and a storage rack. One of the decision problems for this system is the scheduling problem of the two lifts. In other words, which lift is going to handle which request and in which order. In this paper, we derive an integrated look-ahead heuristic based on enumeration to simultaneously assign a set of pre-defined requests to the lifts and to schedule the lifts. As main performance measure we use the total time required to serve all requests.
Class-Based Storage with a Finite Number of Items Yugang Yu and René B.M. de Koster ABC class-based storage is widely studied in literature and applied practice. It divides all stored items into a limited number of classes according to their demand rates (turnover per unit time). Classes of items with higher turnovers are stored in a region closer to the warehouse depot. In literature, it is commonly shown that the use of more storage classes leads to shorter travel time for storing and retrieving items. A basic assumption in this literature commonly is that the required storage space of items equals their average inventory levels, which is right if an infinite number of items are stored in each storage region. However, if a finite number of items are stored in the warehouse, more storage classes need more space to store the items: more classes lead to fewer items stored per class, which have less opportunity to share space with other items. This paper revisits ABC class-based storage by relaxing the common assumption that the total required storage space of all items is independent of the number of classes. We develop a travel time model and use it for optimizing the number and the boundaries of classes. Our numerical results illustrate that a small number of classes is optimal.
Compring Transport Policies in a Full-Scale 300MM Wafer Manufacturing Facility Claude Yugma, Jean-Etienne Kiba, Stéphane Dauzère-Péres and Gilles Lamiable Research in semiconductor manufacturing ideally wants to determine the “best” transport policy to ensure continuous production. Determining such a policy is difficult because it depends on many factors such as the layout, the product types, the equipment, etc. Most of the transport policies found in the literature combine dispatching policies (scheduling of transport requests) and routing policies (selection of the path to move from one point to another). This paper investigates a policy called "minimum service" which consists in keeping a minimum number of available vehicles in bays, so that they can quickly answer transport requests and empty travel times can be minimized. This paper aims at comparing, through experimental tests on actual instances of a real semiconductor manufacturing facility, two types of transport policies in terms of cycle time, throughput and Carrier Exchange Time. Moreover, the behavior of the “minimum service” policy is studied when the number of vehicles and the number of starting lots are varied. The results show that the “minimum service” policy is in general more effective than a classical policy, but that its key parameters must carefully be determined.