Particle Swarm Optimization for the Assignment of Facilities to Locations

Abstract

This chapter describes a new heuristic approach, for minimizing discrete space functions. The new heuristic, particle swarm optimization is applied to the quadratic assignment problem. It is observed from experimentation that the particle swarm optimization approach delivers competitive solutions when compared to ant system, ant system with non-deterministic hill climbing, simulated annealing, tabu search, genetic algorithm, evolutionary strategy, and sampling & clustering for the quadratic assignment problem. By comparing results from the particle swarm optimization and the results of these other best-known heuristics, it will be demonstrated that the particle swarm optimization method converges as much as best-known heuristics for the QAP. The new method requires few control variables, is versatile, is robust, easy to implement and easy to use.