Multi-Objective Particle Swarm Optimizers: An Experimental Comparison Juan J. Durillo 1 , Jos´ e Garc´ ıa-Nieto 1 , Antonio J. Nebro 1 Carlos A. Coello Coello 2 , Francisco Luna 1 , and Enrique Alba 1 1 Department of Computer Science, University of M´alaga (Spain) {durillo, jnieto,antonio,flv,eat}@lcc.uma.es 2 Department of Computer Science, CINVESTAV-IPN, Mexico ccoello@cs.cinvestav.mx Abstract. Particle Swarm Optimization (PSO) has received increased attention in the optimization research community since its first appear- ance. Regarding multi-objective optimization, a considerable number of algorithms based on Multi-Objective Particle Swarm Optimizers (MOP- SOs) can be found in the specialized literature. Unfortunately, no exper- imental comparisons have been made in order to clarify which version of MOPSO shows the best performance. In this paper, we use a benchmark composed of three well-known problem families (ZDT, DTLZ, and WFG) with the aim of analyzing the search capabilities of six representative state-of-the-art MOPSOs, namely, NSPSO, SigmaMOPSO, OMOPSO, AMOPSO, MOPSOpd, and CLMOPSO. We additionally propose a new MOPSO algorithm, called SMPSO, characterized by including a velocity constraint mechanism, obtaining promising results where the rest per- form inadequately. Key words: Particle Swarm Optimization, Multi-Objective Optimiza- tion, Comparative Study 1 Introduction The relative simplicity and competitive performance of the Particle Swam Op- timization (PSO) [11] algorithm as a single-objective optimizer have favored the use of this bio-inspired technique when dealing with many real-word opti- mization problems [17]. A considerable number of these optimization problems requires to optimize more than one objective at the same time which are in conflict with respect to each other. These properties, along with the fact that PSO is a population-based metaheuristic, have made it a natural candidate to be extended for multi-objective optimization. Since the first proposed Multi- Objective Particle Swarm Optimizer (MOPSO) developed by Moore and Chap- man in 1999 [15], more than thirty different MOPSOs have been reported in the specialized literature. Reyes and Coello [17] carried out a survey of the existing MOPSOs, providing a complete taxonomy of such algorithms. In that work, the authors considered as the main features of all existing MOPSOs the following