Archiving with Guaranteed Convergence and Diversity in Multi-Objective Optimization
Marco Laumanns, Lothar Thiele, Eckart Zitzler, Kalyanmoy Deb · 2002
Over the past few years, the research on evolu-tionary algorithms has demonstrated their niche in solving multi-objective optimization prob-lems, where the goal is to find a number of Pareto-optimal solutions in a single simulation run. However, none of the multi-objective evo-lutionary algorithms (MOEAs) has a proof of convergence to the true Pareto-optimal solutions with a wide diversity among the solutions. In this paper we discuss why a number of earlier MOEAs do not have such properties. A new archiving strategy is proposed that maintains a subset of the generated solutions. It guaran-tees convergence and diversity according to well-defined criteria, i.e. -dominance and -Pareto optimality. 1