Abstract
Capacitated facility location is a general and important issue which needs a quite profound knowledge for long-term planning, and the problem has been widely researched in various industries to determine the facility location and related transportation strategy between facilities with certain capacity. To co-operate an industry, a supply network constructed by multi-stage: suppliers, plants, distribution centers, customers in which the location has decisive influence and should be considered simultaneously. Multiple objectives involving quantitative and qualitative factors are also pursued for more comprehensive decision making when constructing multiple facilities. Classical multi-objective programming relies on predetermined preference by decision marker and provide a single solution. However, in multi-objective problem, there is a Pareto set of non-dominated solutions and both objective should be achieved simultaneously without sacrificing anyone. In this research, a new multi-objective evolutionary algorithm first integrating non-dominated sorting concept in Simplified swarm optimization is proposed to solve multi-objective and multi-stage capacitated facility location problem and provide decision makers a Pareto set of compromise solutions. Compare to possibilistic linear programming, Non-dominated sorting Genetic algorithm II (NSGAII), Non-dominated sorting particle swarm optimizer (NSPSO) and Multi-objective particle swarm optimization (MOPSO), numerical results show that the proposed approach can successfully obtain a perfect Pareto set in terms of quality and diversity, even regarded as a competitive approach in multi-objective problem.