Portfolio optimization is an important problem in finance. Its goal is to discover an efficient frontier which shows highest expected return on each level of portfolio variance. The problem has multiple objectives and its search space is large. Multi-objective particle swarm optimization is a multi-objective optimization method, developed from particle swarm optimization by applying non-dominated sorting and crowding distance. This research proposes a portfolio optimization technique based on multi-objective particle swarm optimization. Two objectives used in the research are the maximization of return and the minimization of portfolio risk. The technique is evaluated using daily stock total return index gross dividends from Stock Exchange of Thailand between year 2006 and 2014. The technique is deployed in unknown trading periods, and the results are compared with standard market benchmarks. The results show that the proposed technique outperforms all the benchmarks.
Viriya Yimying, National Institute of Development Administration, Thailand
Ohm Sornil, National Institute of Development Administration, Thailand
Stream: New Realities through Artificial Intelligence
This paper is part of the ACTIS2015 Conference Proceedings (View)
View / Download the full paper in a new tab/window