A Computational Model for Warehouse Analysis and Design Karthik N. Karathur, T. Govindaraj, Douglas A. Bodner, Leon F. McGinnis Keck Virtual Factory Lab – School of Industrial and Systems Engineering Georgia Institute of Technology Atlanta, GA 30332-0205, USA Abstract Over the years, researchers in the field of warehousing systems have developed many analytical models and methodologies. Yet, few have found use in the hands of design practitioners as an aid for design of warehousing systems. Instead, designers prefer to rely on their own expertise. A comprehensive design tool is being created based on a process model [1] that was formulated through detailed ethnographic study [2] of expert warehouse designers as they worked through design projects. The tool is intended to follow the pattern of the cognitive processes of experts as they try to arrive at a design solution, and serve as a design aid to the experts. This is especially important because it attempts to bridge the gap between the theoretical and practical realms of application. In this paper, the proposed conceptual and logical structure of the tool is described, work on which has commenced. Keywords: Warehousing system, ethnographic studies, process model, computational framework, design. 1. Introduction Escalating costs of labor and raw materials and the need for lower lead times have forced companies to rethink their supply chain strategies. As a crucial element of the supply chain, warehousing systems have been squeezed into making work practices better to increase service levels and decrease operating costs. The economic interest in such improvements has spurred companies to back research efforts that aim to develop new methodologies to make warehousing operations more efficient. Research has generally focused on specific aspects of warehouse design such as storage policies [3][4], order picking [5][6] and facilities layout and allocation [7][8]. Being too case-specific, they do not provide perspectives on warehouse design as a whole. Hence, they have not been too helpful for practical use in the actual process of warehouse design. There is no formalized methodology that integrates tools that solve specific problems into a coherent framework for design. Without such a framework, designers are forced to develop their own heuristics to solve design problems. This has caused a disconnect between the theoretical and practical realms of application. The authors observed that designers use little more than spreadsheets such as MSExcel. At the most, a few customized macros in the spreadsheet are used to accomplish data profiling. Rouwenhorst et al. [9] and van den Berg [10] present comprehensive literature reviews to underscore this disconnect. They support the benefits of current research while also stating that most papers analytically investigate well-defined problems in isolation of other warehouse issues. They emphasize the need for combined models and techniques to develop a complete warehouse design approach. If a comprehensive computational model could be developed that integrates the most relevant and practical of these tools in a structure that closely resembles the actual design process as accomplished by experts, it could serve as a useful design aid. If successful, it could prove to be an important asset both for practical use and for validating theoretical tools and methodologies. In the section that follows, the process model on which the computational model is based is discussed. A review of a few methodologies that could be used to approximate the decision-making logic detailed by the process model is presented in section 3. This section includes the design of the decision-making framework on which the computational model is based. The working of the computational model and its software implementation are discussed in sections 4 and 5 respectively. In section 6, a brief summary of the paper and the future course of action are presented. 2. Process Model As stated earlier, warehouse designers use their expertise to arrive at a reasonable solution given the limited amount of time. They do not have the time to consult the various research materials that are available to develop the “best” possible solution. Designers strive to satisfice [11] (reach satisfactory solutions using