A multi-attribute layout design problem

Abstract

Plant layout procedures available today consider either the quantitative attribute or the qualitative attribute of the layout problem. These procedures are based on maximizing or minimizing one objective function. Some of the procedures that address the multi-attribute nature of the plant layout problem are based on assigning weights to objective functions and then solving the problem by heuristic methods. Exact methods guarantee optimal solution but are computationally intensive for large sized problems. Heuristic methods do not guarantee optimal solution, but are computationally less intensive and relatively faster than the exact methods.

This thesis suggests a procedure for solving the multi-attribute plant layout problem. The procedure is based on formulating the plant layout problem as a Quadratic Assignment Problem. Solution to the formulation is achieved by solving it by a combination of exact and heuristic methods. This solution procedure gives better results than just the heuristic method and is not computationally intensive.

A computer implementation of the proposed procedure is developed in "C" programming language. A detailed experimentation is conducted to study the performance of the proposed procedure as compared to a well known procedure called Blocplan. The layouts generated by the proposed procedure are found to be much better than those generated by Blocplan.

In the literature there is no discussion on the criticality of a layout to the change in flow related input data. An analysis is performed to study the affect of change in flow related input data on the final layout and a measure is developed for determining the criticality of a given layout to the change in flow related input data for the case of equal size departments.