A Particle Swarm Optimization Algorithm for Multiobjective Design Problems

A new Particle Swarm Optimization algorithm to solve multiobjective optimization problems is presented. After an initial search, the swarm is subdivided in a number of sub-swarms depending on the distance of each individual from the Pareto frontier. Effectiveness and efficiency of the proposed approach are preliminary investigated by solving a set of well- known test problems, and comparing the results with other evolutionary algorithms. Finally its usefulness is demonstrated in the multiobjective design of a containership.