Accession Number : ADA329129
Title : An Empirical Comparison of Tabu Search, Simulated Annealing, and Genetic Algorithms for Facilities Locations Problems.
Descriptive Note : Doctoral thesis,
Corporate Author : AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH
Personal Author(s) : Arostegui, Marvin A., Jr
PDF Url : ADA329129
Report Date : 02 SEP 1997
Pagination or Media Count : 478
Abstract : Operations managers are typically faced with the need to find good solutions to difficult problems. Such problems include job scheduling, assembly line balancing, process layout, project scheduling, and facilities locations. Although optimal solutions are preferable, the combinatorial nature of these problems means that in many cases problems found in practical applications cannot be solved to optimality within reasonable resources. In these cases, operations managers turns to heuristics. Since the early 1980s, much interest has been devoted to the development and application of three general heuristic algorithms: tabu search, simulated annealing, and genetic algorithms. Each of them specifies a strategy for searching the solution space of a problem looking for "good" local optima. From a practical point of view, we would like to know if any of these methods is indeed better than the other two. In this research study we conduct an empirical comparison of these three heuristic algorithms using three variants of the facilities location problem: capacitated (CFLP), multiple-periods (MP-FLP), and multiple-commodities (MC-FLP). The selection of three different problem structures allowed us to explore the behavior of the heuristics under different circumstances and constraints. Furthermore, none of the heuristics have been previously applied to these problems.
Descriptors : *ALGORITHMS, *ANNEALING, *HEURISTIC METHODS, SIMULATION, POSITION(LOCATION), COMPUTATIONS, JOBS, PARAMETERS, COMPARISON, FACILITIES, STRUCTURES, VARIATIONS, SOLUTIONS(GENERAL), SCHEDULING, MANAGEMENT PERSONNEL, OPERATIONS RESEARCH, ASSEMBLY LINES.
Subject Categories : Numerical Mathematics
Distribution Statement : APPROVED FOR PUBLIC RELEASE