Title: Portfolio optimisation based on minimum total risk acceptance level and its components using improved genetic algorithm

Authors: Sahar Mojaver Rostami

Addresses: Young Researchers and Elite Club, Behshahr Branch, Islamic Azad University, Behshahr, Iran

Abstract: Risk and return are two factors that have always been considered in investment field. Along with the emergence of portfolio optimisation models, with the Markowitz model as the most important one, the need of understanding the methods of solving these models became important as well. One of the most important meta-heuristic methods for portfolio optimisation models is improved genetic algorithm (GA). One goal of this research is to study the efficacy of this algorithm in portfolio optimisation. For this purpose, we once compute the efficient frontier and compare it with efficient frontiers obtained from exact method. To do so, 25 companies from Tehran Stock Exchange companies are selected for this purpose. Calculations are done in Matlab7.6 software. Result show that efficient frontier obtained by GA is the same as that of exact method's, which indicates the high efficiency of GA in portfolio optimisation. Also, this research shows that stock diversification in portfolios with an unsystematic risk function is better than in those with a systematic risk function embedded. The proposed algorithm can solve the portfolio optimisation based on minimum total risk acceptance level and its components by using genetic algorithm. An improved GA is proposed for this purpose.

Keywords: portfolio optimisation; risk acceptance level; Markowitz model; genetic algorithm; systematic risk; unsystematic risk.

DOI: 10.1504/IJCSYSE.2018.095587

International Journal of Computational Systems Engineering, 2018 Vol.4 No.4, pp.248 - 263

Received: 23 Dec 2017
Accepted: 12 Feb 2018
Published online: 11 Oct 2018 *

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