Multi-period M-SV Portfolio Optimization Model with Transaction Costs and Risk Preference

摘要

Portfolio problem is the most popular problem in economics. On the basis of the studies for portfolio optimization problem, we formulate multi-period mean semi-variance optimization model focusing on the transaction costs and risk preference using a weighting parameter. Because of these complex constraints, traditional optimization algorithms can't solve the problem efficiently, we introduce an evolutionary algorithm: Imperialist Competitive Algorithm (ICA). The test data is from the Shanghai Stock Exchange 50 Index (the SSE SO Index). Individual stock's historical daily returns are selected from 1 January 2009 to 3 April 2009. The results show that imperialist competitive algorithm optimization approach is successful in portfolio optimization.