Title: Solving a portfolio optimisation problem via heuristic algorithms

Authors: Xin-Yao Song; Can Cui; Xiao-Shuang Chen; Qi Kang

Addresses: Department of Control Science and Engineering, Tongji University, Shanghai 201804, China ' Department of Information Management, Shanghai Normal University, Shanghai 200234, China ' Department of Control Science and Engineering, Tongji University, Shanghai 201804, China ' Department of Control Science and Engineering, Tongji University, Shanghai 201804, China

Abstract: Portfolio investment has become a fashionable investment with the advantage of risk dispersion. In this work, we examine the ability of three heuristic algorithms to distribute capital associated with the standard mean-variance portfolio optimisation with different risks as constraints. Penalty function is used to deal with the constraint conditions. The three methods considered are Differential Evolution (DE), Comprehensive Learning Particle Swarm Optimiser (CLPSO) and Covariance Matrix Adaption Evolution Strategy (CMA-ES). The results demonstrate that CMA-ES is superior in solving investment portfolio problems in comparison to the other two heuristic algorithms.

Keywords: portfolio optimisation; constrained optimisation; heuristics; portfolio investment; penalty function; differential evolution; comprehensive learning PSO; particle swarm optimisation; CLPSO; covariance matrix adaption evolution strategy; CMA-ES.

DOI: 10.1504/IJWMC.2015.072572

International Journal of Wireless and Mobile Computing, 2015 Vol.9 No.2, pp.125 - 132

Received: 18 Mar 2015
Accepted: 25 Apr 2015

Published online: 19 Oct 2015 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article