• SOLVING CONSTRAINED MULTIOBJECTIVE OPTIMIZATION USING RP-TR/PSO

Abstract

In this paper, a hybrid approach combining trust region (TR) algorithm and particle swarm optimization (PSO) is proposed for solving constrained multiobjective optimization problems (MOOPs). The proposed approach integrates the merits of both TR and PSO. the constrained MOOP is handled by reference point (RP) Interactive Approach and some of the points in the search space are generated. Secondly, For each point the TR algorithm is used to obtain a point on the Pareto frontier. Finally, All the points that have been obtained on the Pareto frontier are used as particles position for PSO; where homogeneous PSO is applied to get all the points on the Pareto frontier. To restrict velocity of the particles and control it, a dynamic constriction factor is presented. The algorithm is coded in MATLAB 7 using 3 GHz PC. Various kinds of multiobjective (MO) benchmark problems have been tested to illustrate the successful result in finding a Pareto-optimal set.