Research on Portfolio Optimization Based on Affinity Propagation and Genetic Algorithm

Abstract

Portfolio optimization refers to the reasonable allocation of assets to achieve the investment objectives. At present, the investment environment depression was due to sluggish economic conditions. For investors, they expect to find a balance between return and risk in a complex investment environment. In order to solve this problem, this paper proposes a portfolio optimization algorithm named Portfolio Optimization based on Affinity propagation and Genetic algorithm, as also called POGA, which based on affinity propagation and genetic algorithm. Firstly, the affinity propagation algorithm is used to construct a candidate set of portfolio based on the correlation analysis of the stock time series. Secondly, using the Sharpe-ratio as the Optimization objective function, the genetic algorithm is used to solve an optimal portfolio strategy with higher-return and lower-risk. Finally, the experimental result of real-world stock data show that a portfolio with higher return and lower risk will be selected.