Solving nonlinear constrained optimization problems using hybrid evolutionary algorithms

Abstract

An optimization problem is the problem of finding the best solution from all feasible solutions. Solving optimization problems can be performed by heuristic algorithms or classical optimization methods. The aim of this article is to introduce a hybrid evolutionary algorithm based on Particle Swarm Optimization(PSO) and Genetic Algorithm(GA). The proposed algorithm consists of hybrid iterations. Each hybrid iteration contains two iterations, Particle Swarm Optimization (PSO) and Genetic Algorithm(GA) iteration. The basic idea is to transmit a population which is performed after the Particle Swarm Optimization (PSO) iterations to be the initial population of the first iteration in Genetic Algorithm (GA), and then continue the rest number of Genetic Algorithm (GA) iterations. The final population of Genetic Algorithm(GA) iteration is used as initial population of the Particle Swarm Optimization(PSO) iteration in the next hybrid iteration. The proposed algorithm is tested on 5 well known test problems. Comparison established against other algorithms proves that the proposed algorithm preserve finding the optimal solution while reduces the function evaluations.