A comparison of multi-objective optimisation metaheuristics on the 2D airfoil design problem

Abstract

In this paper different variants of Multi-Objective Particle Swarm Optimisation (MOPSO) algorithm are investigated, mainly focused on the swarm topology, to optimise the well-known 2D airfoil design problem. The topologies used are global best, local best, wheel, and Von Neumann. The results are compared to the Non-dominated Sorting Genetic Algorithm (NSGA-II) and Multi-objective Tabu Search (MOTS) algorithm, and show that the attainment surfaces achieved by some of the MOPSO variants completely dominate that of NSGA-II. In general, the MOPSO algorithms significantly improve diversity of solutions compared to MOTS as well. The MOPSO algorithm proves its ability to exploit promising solutions in the presence of a large number of infeasible solutions, making it well-suited to problems of this nature.