PSO Application in CVAR Model

Abstract

In order to solve mean variance model with the conditional value at risk (CVaR), an improvement PSO with the generalized learning and the hybrid mutation of dynamic cauchy and the normal cloud model (PSOHM) is proposed to increase the diversity of the population. In PSOHM, to enhance the ability of the population, the introduction of a generalized learning strategy is introduced to enhance flying to the optimal solution for the whole swarm, and according to swarm performance, two different mutation is stimulated to produce the new individual to guide the population flying better. In benchmark function test, the result shows that PSOHM has better performance results. In the portfolio optimization model of CVaR, PSOHM has a better results compared with other algorithms.