基于遗传–蚁群算法的证券组合投资优化研究  

Research on Protfolio Investment Optimization Based on Genetic-Ant Colony Algorithm

作　　者：王慧颖 衣梦涵 刘昊 

机构地区：[1]中央财经大学统计与数学学院,北京 [2]东北农业大学理学院,黑龙江哈尔滨 [3]中央财经大学管理科学与工程学院,北京

出　　处：《建模与仿真》2016年第4期161-169,共9页Modeling and Simulation

基　　金：感谢国家自然科学基金项目(11301558);2014年度中央财经大学重大科研课题培育项目(基础理论类)的资助。

摘　　要：基于Markowitz资产组合理论,综合考虑证券投资的风险和收益,建立证券组合投资的多目标规划模型,融合遗传算法和蚁群算法应用于上述模型的求解。具体地,将具有快速全局搜索能力的遗传算法产生的问题初始解转化为蚁群算法的初始信息分布,再对蚁群进行遗传操作,最后利用蚁群算法的并行性、正反馈机制、求解效率高的特征寻求最优解。实验结果表明:上述两种算法的融合在求解质量和效率上均优于单独的遗传算法或蚁群算法。Based on the Markowitz portfolio theory, the multi-objective programming model of portfolio investment is established when considering the risk and return of portfolio investment. The genetic algorithm and ant colony algorithm are combined and applied to solve the above model. In detail, the initial solution of problems which is generated by the genetic algorithm with fast global searching ability is transformed into the initial information distribution of the ant colony algorithm. And then we carry on the genetic operation to the ant colony. Finally, we use the ant colony algorithm parallelism, positive feedback mechanism and the solution efficiency high characteristic to seek the optimal solution. Experiments show that: the fusion of the two algorithms has better behaviors on the quality and efficiency than separate genetic algorithm or ant colony algorithm.

关 键 词：证券组合投资 多目标规划 遗传算法 蚁群算法 算法融合 

分 类 号：F2[经济管理—国民经济]