针对资产数目和投资资金比例受约束的投资组合选择这一NP难问题,基于混沌搜索、粒子群优化和引力搜索算法提出了一种新的混合元启发式搜索算法。该算法能很好地平衡开发能力和勘探能力,有效抑制了算法早熟收敛现象。标准测试函数的测试结果表明混合算法与标准的粒子群优化和引力搜索算法相比具有更好的寻优效率;实证分析进一步对混合算法与遗传算法及粒子群优化算法在求解这类投资组合选择问题的性能进行了比较。数值结果表明,混合算法在搜索具有高预期回报的非支配投资组合方面表现更好,取得了更为满意的结果。
The portfolio selection problem with constraints of asset number and investment proportion is a NP-hard problem. This paper proposed a new hybrid meta-heuristic algorithm with the combination of chaotic search, particle swarm optimiza- tion, and gravitational search algorithm for solving this problem. The hybrid algorithm efficiently balanced the ability of exploi- tation and the ability of exploration, and adaptively avoided the stagnancy of population and increased the speed of conver- gence. Some benchmark test functions were used to compare the hybrid algorithm with both the standard PSO and GSA algo- rithms in optimization efficiency. The results show the hybrid algorithm is better than them. The performance of the proposed algorithm was also compared with GA and PSO by an empirical analysis for portfolio selection with cardinality constraints. The numerical results demonstrate that the proposed hybrid algorithm can achieve satisfactory results and perform well in searching non dominated portfolios with high expected returns.