Improved particle swarm optimization and application to portfolio selection

Abstract

Particle swarm optimization (PSO) is a population-based stochastic optimization technique, inspired by the social behavior of birds (flocking) or fish (schooling), which is applied to various problems in the optimization of nonlinear systems. The inertia weights approach (IWA) and the constriction factor approach (CFA) are improved methods in PSO. The IWA searches the problem space globally in the early steps, and finally searches locally near the optimal solution. CFA is a method that introduces a new parameter into velocity update equation. This paper proposes a combination of IWA and CFA (the Inertia Weights Constriction Factor Approach: IWCFA), and PSOrank, whose objective is the ranking of individuals in the population. These two proposed methods are applied to function optimizations and to the portfolio selection problem, which is a typical mathematical problem in securities finance. The results show that the original PSO finds better solutions than the GA, and the proposed method finds better solutions than the original PSO. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(3): 13–25, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20263