Study on the Efficient Frontier in Portfolio Selection by Using Particle Swarm Optimization

摘要

In this paper, an approach is presented to compute the efficient frontier for portfolio optimization based on particle swarm optimization (PSO). A generalization of the standard Markowitz mean-variance model which includes transaction costs and floor and ceiling constraints is considered. Due to these complex practical constrains exact algorithms fail to work efficiently, so the use of heuristic algorithms in this case is imperative. At last, some experimental results is presented and the efficient frontier under different constrains is compared. Simulation results show that the PSO algorithm converges quickly with consistent performance, which make it suitable for creating efficient frontier for much larger number of assets.