Behavioral approach in multi-period portfolio optimization using genetic algorithm

This paper discusses a multi-period portfolio optimization problem by considering a conditional value-at-risk (CVaR) constraint Based on prospect theory, which considers the loss-averse utility, the transaction cost and the lower bound and upper bound investment in each asset. A genetic algorithm is proposed to solve the portfolio model. The results based on the average optimal ultimate wealth and Sharp ratio criteria showed that loss-averse investors tend to concentrate most of their wealth and perform better than rational investors. The impact of CVaR on investment performance was identified. When the market falls, investors with higher risk aversion avoid extreme losses and obtain more gains.