This paper concerns the problem of portfolio optimization in dynamic environments using multi-objective evolutionary algorithms. Financial markets are characterized by volatility and uncertainty making portfolio optimization a challenging task. A novel multi-objective genetic programming algorithm is proposed, which is a memory enhanced version of the standard SPEAII algorithm. The proposed algorithm employs an explicit memory to store a number of non-dominated solutions. These solutions are reused in the later stages for adapting to the changing environments. A stock ranking based trading simulation is used for fitness evaluation and a probabilistic metric is employed to choose a solution from the Pareto Front. The experiments are performed on the constituent stocks of the S&P BSE FMCG Index. The results, evaluated using RMSE, MEA and cumulative returns, are very promising.
CITATION STYLE
Shah, P., & Shah, S. (2017). Portfolio optimization in dynamic environments using MemSPEAII. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10682 LNAI, pp. 424–436). Springer Verlag. https://doi.org/10.1007/978-3-319-71928-3_40
Mendeley helps you to discover research relevant for your work.