基于收益权重的多期投资组合优化模型

Multi-period Portfolio Optimization Model Based on Return Weight

为解决经典均值-方差模型对收益率的假设不能完全适应于实际证券市场的问题,引入收益权重计算每个时间段收益占总时间段收益的比重。考虑模糊环境下多重摩擦因素,提出了一种可能性均值-下半方差-熵多周期投资组合优化模型。利用外点罚函数法将其转化为无约束优化问题。在标准的遗传算法中引入混沌映射改变种群提取过程,扩大种群多样性以便寻到更优个体。数值实验和仿真结果表明建立的模型是合理的,改进的算法是有效可行的。

To solve the problem that the return assumption in the classical mean-semi-variance-entropy model cannot be fully adapted to the actual stock market,return weights are introuced to calculate the proportion of the return of each time period to the total time period return.A multi-period portfolio optimization model with possibility mean-semi-variance-entropy is proposed considering the factors of multiple frictions in a fuzzy environment.Then,the model is transformed into an unconstrained optimization problem using the external penalty function method,and an improved genetic algorithm is used to solve the problem,which introduce the chaotic mapping to change the population extraction process of the standard genetic algorithm to expand the population diversity and find better individuals.Numerical experiments and simulation results show that the established model is reasonable,and the improved algorithm is effective and feasible.