Multi-Objective Particle Swarm Optimization Based on Fuzzy Optimality

In order to overcome the limitations of the Pareto optimality in solving multi-objective optimization problems, a new optimality definition, fuzzy optimality is proposed, which considered both of the numbers of improved objectives and the extent of the improvements. Then, the fuzzy optimality-based multi-objective particle swarm optimization algorithm is presented. It inherits the basic structure of the particle swarm optimization and evaluates the particles by the fuzzy optimality. The numerical experiments are carried out on 6 representative test functions, and the results show that the proposed fuzzy optimality based multi-objective particle swarm optimization algorithm shows better performance on aspects of quality of solutions, robustness, and computational complexity, compared with the results of the NSGA-II and MOPSO. Finally, the efficacy and practicality of the proposed approach are validated in the APU fuel consumption and emissions multi-objective optimization problem.