A Novel Multi-objective Optimization Algorithm Based on Artificial Immune System

Abstract

The traditional evolutionary algorithm (EA) for solving the multi-objective optimization problem (MOP) is difficult to accelerate convergence and keep the diversity of the achieved Pareto optimal solutions. A novel EA, i.e., Immune Multi-objective Optimization Algorithm (IMOA), is proposed to solve the MOP in this paper. The special evolutional mechanism of the artificial immune system (AIS) prevents the prematurity and quickens the convergence of optimization. The method combined by the random weighted method and the adaptive weighted method guarantee the acquired solutions to distribute on the Pareto front uniformly and widely. An external set for storing the Pareto optimal solutions is built up and updated by a novel approach. By graphical presentation and examination of selected performance metrics on two difficult test functions, the proposed IMOA is found to outperform four other algorithms in terms of finding a diverse set of solutions and converging near the true Pareto front.