Co-operative Co-evolution Based Hybridization of Differential Evolution and Particle Swarm Optimization Algorithms in Distributed Environment

Evolutionary computing algorithms play a great role in solving real time optimization problems. One of the evolutionary computing algorithm is Particle Swarm Optimization algorithm (PSO). The aim of this paper is to propose a model to improve the performance of PSO algorithm. Hybrid models of Particle Swarm Optimization (PSO) algorithm and Differential Evolution (DE) has already proved to be one of the better approaches for solving real world complex, dynamic and multimodal optimization problems. But these models hybridize PSO and DE to form a new serial algorithm. In these serial hybridization models, we are losing the originality of both DE and PSO algorithms since the structure of both the algorithms is being modified to get the hybridized PSO and DE algorithm. In this paper, we develop a model for PSO in distributed environment with improved performance in terms of speed and accuracy. The proposed model is a hybridized distributed mixing of DE and PSO (dm-DEPSO) which improves the performance of PSO algorithm. In this model, algorithms are implemented in a cluster environment to perform co-operative co-evolution. Better solutions are migrated from one node to another in the cluster environment. Co-operative co-evolving model shows better performance in terms of speed and accuracy. The algorithm is applied to a set of eight benchmarking functions and their performance are compared by mean of objective function values, standard deviation of objective function values, success rate, probability of convergence and quality measure.