Coevolutionary Comprehensive Learning Particle Swarm Optimizer

Abstract

In this paper, a Coevolutionary Comprehensive Learning Particle Optimizer (Co-CLPSO) is proposed for solving constrained real-parameter optimization problems. In this novel algorithm, a coevolutionary schedule and a novel constraint-handling mechanism are employed. Two swarms with different thresholds are constructed and they exchange information in the evolution process. Different with the existing constraints handling methods, the particles are adaptively assigned to explore different constraints according to their difficulties. These new mechanisms are combined in Comprehensive Learning Particle Swarm Optimizer (CLPSO) and Sequential Quadratic Programming (SQP) method is combined to improve its local search ability. The performance of the proposed Co-CLPSO on the set of benchmark functions provided by CEC2010 [1] is reported.