Multi-Objective Particle Swarm Optimization Algorithm for Unequal-Area Dynamic Facility Layout Problem

This article studies the unequal-area dynamic facility layout problem (UA-DFLP) with two objectives derived from the production and manufacturing system of the manufacturing industry, which is an NP-hard problem with wide applications. A multi-objective particle swarm optimization algorithm based on objective space segmentation (MOPSO-OSS) which divides the objective space of the problem into some cells is proposed for the UA-DFLP. The algorithm constructs a new fitness value function based on the superiority and density of particles by controlling the neighborhood topology structure and selects the historical and global optimal positions of particles based on the fitness value. In addition, a heuristic mutation strategy and an adaptive gradient method are proposed to address the problem of embedding between facilities. The experimental results, running on two classic examples and a practical production application, demonstrate the efficacy of the proposed MOPSO-OSS algorithm in tackling the UA-DFLP.