Hybrid Artificial Immune Algorithm and Particle Swarm Optimization for Solving Unconstrained Global Optimization Problems

Abstract

This work presents a meta-heuristic approach that integrates an artificial immune algorithm and a particle swarm optimization (AIA-PSO) method to solve unconstrained global optimization (UGO) problems. Using an external AIA, the parameter settings of an internal PSO algorithm are optimized. These include the constriction coefficient, cognitive parameter and social parameter. The internal PSO algorithm is used to solve benchmark UGO problems. Furthermore, this work compares the numerical results obtained using the proposed AIA-PSO algorithm with those obtained using published Nelder¡VMead simplex search method ¡V PSO (NM-PSO), particle swarm ant colony optimization (PSACO), genetic algorithm-PSO (GA-PSO), continuous hybrid algorithm (CHA) and continuous tabu simplex search (CTSS). Experimental results indicate that the proposed AIA-PSO algorithm converges to a global optimum solution to each UGO problem. Furthermore, the optimum parameter settings of the internal PSO algorithm can be obtained using the external AIA. Moreover, the proposed AIA-PSO algorithm outperforms those of some published NM-PSO, GA-PSO, CHA and CTSS for each UGO problem. Therefore, the proposed AIA-PSO algorithm is a highly promising alternative stochastic global optimization method for solving UGO problems.