About | Research | Events | People | Reports | Alumni | Contact | Home
Workshop Stochastic models for warehousing systems October 29-30, 2009 Programme & Abstracts Thursday October 29, 2009
Friday October 30, 2009
Ricky Andriansyah, Eindhoven University of Technology Simulation Model of a Single-Server Order Picking Workstation using Aggregate Process Times We propose a simulation modeling approach based on aggregate process times for the performance analysis of order picking workstations in automated warehouses with first-in-first-out processing of orders. The aggregate process time distribution is calculated from tote arrival and departure times. We refer to the aggregate process time as the effective process time. We distinguish between the effective process time distribution for the first tote of an order and that of the remaining totes of an order. These two distributions are used in an aggregate model to predict tote and order flow times. Results from a test case show that the aggregate model accurately predicts the mean and variability of tote and order flow times. The effect of the order size distribution on the flow time prediction accuracy is also investigated. Kai Furmans, Universität Karlsruhe Blocking phenomena in Order-picking Systems Productivity in order picking systems suffers, when the persons in the picking system try to access the same space - either walking or picking. We show, how the loss in throughput through blocking can be approximately predicted by modeling the problem as a closed queuing network with limited buffers. We also show, how the methods of Akyldiz and Marie can be combined and enhanced in order to calculate approximate performance figures. Jérémie Gallien, MIT Sloan School of Management To Wave Or Not To Wave? Order Release Policies for Warehouses with an Automated Sorter Abstract: Wave-based release policies are prevalent in warehouses with an automated sorter, and take different forms depending on how much waves overlap and whether the sorter is split. Waveless release is emerging as an alternative policy adopted by an increasing number of firms. While that new policy presents several advantages relative to waves, it also involves the possibility of gridlock at the sorter. In collaboration with a large US online retailer and using an extensive dataset of detailed flow information, we first develop a model with validated predictive accuracy for its warehouses operating under a waveless release policy. We then use that model to compute operational guidelines for dynamically controlling the main parameter of its waveless policy, with the goal of maximizing throughput while keeping the risk of gridlock under a specified threshold. Secondly, we leverage that model and dataset to perform through simulation a performance comparison of wave-based and waveless policies in this context. Our waveless policy yields larger or equal throughput than the best performing wave-based policy with a lower gridlock probability in all scenarios considered. Waveless release policies thus appear to merit very serious consideration by practitioners. Facilities using a non-overlapping wave policy should also consider overlapping waves or a split sorter policy. Presentation Yeming (Yale) Gong, Erasmus University Rotterdam Using stochastic networks to improve the flexibility of warehouses operations in public storage Public storage is a booming industry. A major question is how to improve the flexibility in warehouses operations to fit changing market segments and accommodate volatile demand in order to maximize revenue. Customers that cannot be accommodated with a space size of their choice can be upgraded to a larger space. We use overflow queue networks to describe this upgrading operations. We solve the models for several real warehouse cases, and our results show for all cases the flexibility of the existing public-storage warehouses can be improved. Presentation 1 - Presentation 2 Roelof Hamberg, ESI Development of new warehousing systems and the use of modeling Every new warehouse (conceived with functionality and technologies which are new or existing) has to be developed from customer needs towards an engineered system. Modeling and analysis supports this development process with different models by focusing at different aspects of candidate systems, such as performance metrics, robustness, and cost. Typically, the available level of system detail is increasing as the development progresses, which also impacts the modeling approach that is being or can be applied. Some samples of industrially used modeling approaches in the domain of warehousing are given, together with their natural habitats involving issues such as knowledgeability, levels of abstraction, spent effort, required flexibility, and standardization of components. Our approach, aimed at improved modeling and analysis in early phases of development, means to fit seamlessly in this industrial environment. Sunderesh S. Heragu, University of Louisville Analytical Models for Design, Planning and Operation of Warehousing Systems In this talk, we cover some of the more important strategic, planning and operational problems encountered in storage and retrieval systems. Problems addressed at the strategic level include determining the amount of storage space to allocate to the forward, reserve and cross-docking areas in a warehouse. They also include choosing between alternate material handling technologies in an automated storage and retrieval system (AS/RS). The problems addressed at the planning level include more detailed configuration of a warehouse and its material handling system. Problems addressed at the operational stage involve real-time control of warehouse operations in response to changes taking place dynamically in the warehouse system, due to disturbances internal and external to the system. Depending upon the nature of these problems, they are solved via mathematical programming, closed queuing network, open queuing network, semi-open queuing network, and intelligent agent modeling approaches. These techniques are applied on problems and data from industry. René de Koster, Rotterdam School of Management, Erasmus University Warehouse assessment in a single tour In this presentation I introduce an assessment method for warehouses based on a single facility tour and some Q&A. The method helps managers and students alike that visit a facility to get more information from tour visits through a simple and rapid assessment form. Since its inception, it has been applied to a number of cases, successfully identifying weak and strong points of the operations. Presentation 1 - Presentation 2 Ananth Krishnamurthy,
University of
Wisconsin-Madison Vehicle Interference Effects in Warehousing Systems with Autonomous Vehicles Warehouses that use autonomous vehicles for storage and retrieval operations within multi-tier racking systems rely on vehicles to provide horizontal the movement within a tier and use lifts to provide vertical movement between tiers. In these systems, vehicle interference in the aisles and cross aisles could significantly decrease system throughput and increase cycle times. Queuing models are proposed to analyze these effects and decomposition methods are used to evaluate the effect of vehicle interference on system performance. The models are validated using detailed simulations. Liqiang Liu, EURANDOM Queueing Network Analysis of Compact Picking Systems In this paper we focus on queueing network modeling and performance evaluation of compact picking systems. Our model is a closed, multi-class queueing network which does not possess a product form solution in general. We develop an efficient aggregation/disaggregation type approximation to compute major performance measures accurately. We conduct extensive numerical experiments to test the goodness of the method. José Antonio Larco Martinelli , Erasmus University Rotterdam Behavioral Goal Setting Models for Operations Management Operations Management (OM) work flow models typically assume that workers are unaffected by external factors. This assumption is contradicted by a major theory within the field of Industrial Psychology, referred to Goal Setting Theory, which posits that setting challenging goals enhances performance. This paper reviews Goal Setting Theory from an OM perspective verifying whether the relationship between goal difficulty and performance is strictly non-decreasing and exploring how workers regulate their work pace under the influence of goals. We address these objectives using a two-folded approach. Firstly, two distinct decision making models are proposed in the tradition of behavioral economics to generate alternative propositions regarding the goal difficulty-performance relationship linked to the work pace regulation of workers. Secondly, we conduct a laboratory experiment to verify the validity of these propositions. The results from the experiment provide general support a strictly increasing relationship between goal difficulty and performance and a stationary work pace. However, we also find that it is setting challenging and realistic goals that induce workers to have a stationary work pace. We discuss possible explanations for these findings as well as their implications for OM models and practice. Marco Melacini, Politecnico di Milano Design of OPS: insights from Italian research Since 2003 Politecnico di Milano has established a permanent Observatory whose aim is to study material handling systems. The Observatory is funded by a number of material handling providers (e.g. Dematic, TGW, Incas group). The methodology adopted consisted in case studies jointly with development of frameworks and analytical models for warehouse design. The presentation will focus on the main results of the research conducted so far by the Observatory. First of all the results of an empirical analysis on the Order Picking System (OPS) application fields, carried out on a sample of over 68 Italian warehouses, will be presented. The analysis has been performed through the Logit model, and has highlighted some key parameters in the OPS selection (e.g. order size, number of items, number of order lines picked per day). Furthermore, the first results of the design of a pick–and--sort system are illustrated, and an analytical model is presented to estimate the picking rate in function of wave length. The model, which has been tested through simulation, includes an algorithm to estimate the expected overlapping of order lines. Ben Montreuil, Laval University Evolution of Logistic Center Conceptualization : Meeting the Physical Internet Challenges First we present the overall concept of a Physical Internet and its justification. Then we identify some of the key impacts on logistic centers, including warehousing systems. We emphasize the need for an evolution of the fundamental conceptualization of such centers in a Physical Internet. Finally we synthesize some of the key research challenges. Kees Jan Roodbergen, Erasmus Universiteit Rotterdam Layout methods for order picking areas
The
efficiency of designs for the order picking area is usually evaluated
through either simulation or closed-form approximative formulas. Simulation
allows for the evaluation of the full range of possible layouts and all
common routing methods, however, a significant effort is needed to create
and run the program and statistical methods are needed to verify the proper
ranking of the alternatives. Closed-form formulas typically provide a fast
means to compare alternatives, but they usually provide only an
approximation and they are typically restricted to certain layouts and
control rules. Maria Vlasiou, Eindhoven University of Technology The contraction principle in two-carousel warehousing models Studies in carousel warehouse systems are usually divided between these studies that study numerically existing paradigms of large-scale systems and the studies that study analytically simplified models involving a single carousel. Few analytic studies exist on multi-carousel systems, as the interdependencies arising in a multi-carousel setting make their analytic study cumbersome. In this talk we make a first attempt to address this issue by presenting an analytic study of a two-carousel system that is useful for fast and accurate numerical computations. In particular, under various picking strategies, we study the sojourn time of a picker serving the two carousels and we conclude that the equation describing his sojourn time satisfies a contraction mapping, which allows for the derivation of the sojourn time distribution through an iterative approach. Bruno van Wijngaarden, VanderLande Dynamic models and strategies to control warehouse processes Vanderlande Industries predominantly uses dynamic simulation
models to support system design. The reason for this is that system
performance optimization strategies are relatively easy to implement in this
type of model. Yugang Yu, Erasmus
Universiteit Rotterdam Does Class-Based Storage Really Reduce Travel Time? Class-based storage is widely claimed to reduce travel time in a warehouse as compared to random storage. We study class-based storage allowing a storage space per product larger than its average inventory level. We develop a model and optimize the number and boundaries of classes to minimize the travel time to store or retrieve products. Examples show a larger number of classes may give a longer travel time. Last updated
11-nov-2009,
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
P.O. Box 513, 5600 MB Eindhoven, The Netherlands |