A Hybrid Particle Swarm Optimization for Dynamic Facility Layout Problem

The dynamic facility layout problem (DFLP) aims to minimize the sum of handling and re-layout costs by devising an individual layout for each distinctiveproduction period. It is a complex combinatorial optimization problem. In this paper we use a hybrid Particle Swarm Optimization (PSO) algorithm to solvethe DFLP. We first employ acoding and decoding technique to efficiently map discrete feasible space of the DFLP to a continuous space, The proposed PSO willfurther use this coding technique to explore the continuous solution space. Afterwards, the algorithmdecodes the solutions to its respective feasible solution inthe discrete feasible space and returns the solutions. PSO combines local search (by self-experience) and globalsearch (by neighboring experience), and possesses high search efficiency .For further enhancement, the proposedPSO is hybridized with a simple and fast simulatedannealing. By reasonably combining these two different search algorithms, we develop an implemented hybridoptimization algorithm. The algorithm is capable of being extended to more general cases although equal area machines and standardized handling equipmentswith identical unit costs are assumed for the time being.Computational results test problems from the literature problems show the efficiency of the proposed algorithm compared to other heuristics.

A Hybrid Particle Swarm Optimization for Dynamic Facility Layout Problem Keywords:

Dynamic facility layout , Particle swarm optimization , F actoradics , Simulated Annealing