Application Of Particle Swarm Optimization To Portfolio Selection

Because Conditional Value-at-Risk(CVa R) can measure the potential risks more fully and has good statistical properties, so it is generally considered a better risk measurement tool than Va R. But the Mean-CVa R model is very difficult to be solved directly, so it is necessary to intelligent algorithms for help. Standard particle swarm algorithm principle(PSO) is simple and less parameters, but with slow convergence rate at later period and low precision. Based on the study the mean-CVa R model, this paper improves particle swarm algorithm, inertia weight was firstly increased and then decreased, for every particle, the inertia weight will be set to 0 when its fitness deteriorated, then the algorithm can improve the efficiency and accuracy.This paper majors in the solution of mean-CVa R model and rate of return and risk as equality aim model, using Standard and improved PSO. The main conclusions are as follows:(1) Improved algorithm is more effective and is able to search for better solutions.(2) Between rate of return and CVa R, probably exists "Pareto optimal" curve.