Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Industrial Engineering


Russell D. Meller

Committee Member

John A. White, Jr.

Second Committee Member

Richard E. Webb

Third Committee Member

Sarah E. Root


Applied sciences, Inventory management, Order picking, Replenishment, Warehouse design


Research in the area of warehouse design is characterized by a myriad of analytical models that address one, typically small and isolated, area of the warehouse. These models, although important in gaining insight into one question of warehouse design, are of limited value when one considers the larger question of overall warehouse design. Thus, research in the area of overall warehouse design typically consists of procedure-driven processes based on qualitative factors and not the quantifiable results of analytical models.

In contrast, practitioners have significant empirical data related to how a design alternative performs in an industry, a company, or a particular warehouse. However, because practitioners lack a means for comparing the performance of competing alternatives over multiple facilities, they may adopt a sub-optimal design for a given facility.

A valuable tool for depicting a design is the functional flow network, where nodes represent the functional areas in the warehouse and arcs connecting the nodes define the product flow between functional areas. We propose a design methodology that employs the use of functional flow networks, as well as analytical models and empirical data for quantifying design performance. First, we develop a complete set of analytical models for a manual, case-picking warehouse, and we use the models to investigate the optimal warehouse shape. Next, we implement the design methodology using the analytical models. We then parameterize the analytical models to create lookup tables to demonstrate the design methodology

using empirical data. We use an example to show that the two methods lead to the same solutions, thus providing a proof-of-concept for using empirical data to design a warehouse. Finally, we present a preliminary search heuristic for designing a manual, case-picking warehouse. The search heuristic is based on warehouse operating characteristics and provides an initial design that can be further analyzed and optimized.

We believe that our design methodology provides two key features that are typically

missing from existing overall warehouse design methodologies: comparing design alternatives through quantifiable output from analytical models and empirical observations, and therefore, considering a broad range of design alternatives.