摘要
本文针对均值-CVaR投资组合优化问题,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,而后基于多维布朗运动,借助Monte Carlo模拟情景生成得到价格路径,进而近似求解均值-CVaR投资组合选择问题,并与线性规划和非参数估计两种求解算法进行比较。模拟和实证算例结果表明,新算法在求解有效性和实用性方面表现更好,取得更为满意的结果。
In this paper , a new hybrid heuristic algorithm is proposed with the combination of chaotic search , particle swarm optimization , and gravitational search algorithm for solving mean-CVaR portfolio selection .Monte Carlo simulation is employed for generating scenario paths based on the multivariate brownian motion and the ap -proximate value of CVaR is computed .The computation usefulness and effectiveness between the proposed meth-od and the linear programming method and the methodology of nonparametric estimation are compared by simula-tion and real world examples .Numerical results show that the performance of the new approach is very good .
出处
《运筹与管理》
CSSCI
CSCD
北大核心
2014年第6期229-235,共7页
Operations Research and Management Science
基金
国家自然科学基金资助项目(11171221)
