Multi-period mean-semi-entropy portfolio management with transaction costs and bankruptcy control

This study investigates a multi-period portfolio management problem under fuzzy settings. For the first time, the newly proposed semi-entropy in the literature is employed as an efficient downside risk measure for risk control in multi-period portfolio optimization. Fuzzy techniques for financial modeling show advantageous performance when future financial market conditions cannot be effectively detected with only historical data. We describe the assert returns by fuzzy variables. Two realistic constraints of transaction costs and bankruptcy events are taken into consideration in our model formulation of a multi-period mean-semi-entropy optimization program. The formulated program is rewritten as a crisp single-objective nonlinear programming by introducing a risk-aversion factor, and final solution to the program is obtained by using genetic algorithm. For the demonstration of computational results, we provide a numerical example with real-life stock data, which illustrates the main modelling concept and the efficiency of genetic algorithm solving method. Comparative analyses over several baseline models show the advantages of adopting fuzzy semi-entropy as an efficient downside risk measure for multi-period portfolio investment optimization.