A Mean-Semi-variance Portfolio Optimization Model with Full Transaction Costs

A mean-semi-variance portfolio optimization model with constraints of cardinality, investment quota, total capital and integer transaction is established, the model can reflect the transaction costs completely. Because the conventional genetic algorithm isn´t easy to find feasible solution of portfolio optimization model, a stochastic repair method of an infeasible solution is proposed, which consists of repairs of cardinality, investment quota and total capital constraints. An improved genetic algorithm based the repair of infeasible solution is developed accordingly, which can improve convergence of conventional genetic algorithm. The results of numerical simulation show that the algorithm can solve the model quickly and efficiently.