Date of Award

1989

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Engineering Science (Interdepartmental Program)

First Advisor

John Henry Wells

Abstract

A Robust Construction Heuristic (ARCH) represents the first attempt to develop a facilities layout design algorithm tailored to the particular solution of the food processing facility layout problem where a fluctuating product mix due to seasonality or changes in consumer demand is typical. The foundation literature for this research comes predominantly from the industrial engineering, management science, and food science fields. Construction heuristics built layouts from scratch and provided poor quality results. Robustness algorithms had no simplifying heuristics but were reported to yield potentially higher quality layouts than traditional algorithms for layout problems with fluctuating product mixes. ARCH was developed as an alternative layout design algorithm based on the parallel philosophies that a robustness consideration would improve the relatively low quality solutions reported for construction heuristics and that a construction approach would allow the robust algorithm to be realistically implemented. ARCH was found to provide layout solutions comparable to or better than improvement heuristics for benchmark problems from the literature. A case study of an existing ham processing line was performed and a layout design for the plant generated by ARCH. In-plant surveys were conducted to develop pertinent input data for ARCH. Three products, whole smoked ham, chunked-and-formed ham, and pork sausage were produced by the plant. Historical seasonal demands were used for the expected fluctuation in levels of demand for each product. Certain assumptions had to be made to allow ARCH to consider layouts with certain flow patterns (U-shaped). The resulting layout designs were rather unorthodox in shape but could be used by the human designer to determine adjacency requirements for the final design.

Pages

148

Share

COinS