Implementation of e-New Local Search based Multiobjective Optimization Algorithm and Multiobjective Co-variance based Artificial Bee Colony Algorithm in Stocks Portfolio Optimization Problem

Abstract

The problem of portfolio optimization is a research topic that is quite widely discussed in the financial sector. The first model in this problem is the mean-variance model that focuses on expected return and risk without considering the constraints contained in the real problem. In this paper, a portfolio optimization model with real constraints which is commonly known as the Mean-Variance Cardinality Constrained Portfolio Optimization (MVCCPO) model is considered. The e-New Local Search based Multi-objective Optimization Algorithm (e-NSLS) and Multi-objective Covariance based Artificial Bee Colony (M -CABC) algorithm are used to solve portfolio optimization problem on datasets involving up to 225 assets. Obtained results are compared with the unconstrained efficient frontier of the corresponding data sets. The numerical simulations state that e-NSLS algorithm gives a better solution than M-CABC, where the solutions produced by e-NSLS are nearer to the corresponding unconstrained efficient frontier than the solutions generated by M-CABC.