DEaf-MOPS/D: An Improved Differential Evolution Algorithm for Solving Complex Multi-objective Portfolio Selection Problems Based on Decomposition
Shouheng Tuo, HONG HE · ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH · 2019
With the high-speed economic development and economic diversification of the world, it is very necessary to develop an effective and efficient portfolio selection method with high precision and robustness.In this study, we first introduce an enhanced multi-objective cardinality constrained mean-variance (CCMV) model, in which the transaction costs and price-earning (P/E) ratio are appended in the model, then an improved differential evolution algorithm with adaptive fine-tune is proposed to solve multi-objective portfolio selection problems based on decomposition (DEaf-MOPS/D).Finally, five simulation experiments on five benchmark datasets (HangSeng, DAX 100, FTSE 100, S&P 100, and Nikkei 225) are employed to investigate the performance of our method.The experimental results indicate that the performance of DEaf-MOPS/D is superior to other compared algorithms, and its runtime is much less than other algorithms, which demonstrate that our method is efficient in solving high-dimension portfolio selection problems.