Complex Portfolio Selection Using Improving Particle Swarm Optimization Approach

Complex portfolio selection problems with realistic constraints have being studied by the researchers and practitioners in the financial and economic field. This paper discussed a class of complex portfolio selection problem with cardinality constraints and bonding constraints, which is NP-hard problem, and difficultly tackled by the conventional methods. A heuristic approach, called as particle swarm optimization (PSO), is presented and improved by combination with Extremal optimization (EO). The hybrid approach (eo-PSO) is tested by the data sets from the emerging stock market of china A as well as the classical approaches of GA and original PSO, the comparisons show that eo-PSO is the most competitive. In addition, These results are also investigated by the risk range Theorem (Xue Deng, &Jun-feng Zhao,2013), indicating that the hybrid approach has superior performance.