An Investigation Into the Effect of More Accurate Measures of Distance on the Detailed Facility Layout Problem

摘要

Process Facility Layout Problem (PFLP) involves the arrangement of a given number of departments within a facility in order to minimize the total cost of traveling among departments. Traditionally, Euclidean or Rectilinear distances are utilizedin determining the solution to PFLP. In this paper we attempt to show that using Realistic distance measures improves the solution accuracy for PFLP. We solve a wide range of problems, varying the problem size, flow dominance and distance congruence, using Realistic, Euclidean and Rectilinear distances. The Realistic distances are determined using Dijkstra's Algorithm. Since PFLP is a combinatorial problem which is NP-Complete, we use a heuristic procedure, Tabu Search, in the solution of the problems. The results show that using Realistic distance measures gives better resutls than using the traditional measures. The difference becomes more apparent as flow dominance increases and as the distnace congruence decreases.