遗传算法模型和二次规划算法在投资组合最优化中的应用
发布时间：2018-12-09 12:42
【摘要】：投资组合最优化问题是一个NP难解问题,通常的方法很难使其达到全局最优。而遗传算法模型作为一种高效并行的全局优化搜索方法,已经被应用到很多领域。本文通过将遗传算法模型引入到证券投资分析领域,对最佳证券组合问题进行优化计算,在收益率保持不变的情况下,计算不同投资比例下的风险系数,并且经过与序列二次规划算法计算结果的对比分析,发现使用遗传算法得出的投资组合的风险系数要小于使用序列二次规划方法得出的投资组合的风险系数,由此得出遗传算法模型在某些情况下优于序列二次规划算法的结论。
[Abstract]:Portfolio optimization problem is a difficult NP problem, and it is difficult to achieve global optimization by common methods. As an efficient parallel global optimization search method, genetic algorithm (GA) has been applied in many fields. In this paper, the genetic algorithm model is introduced into the field of securities investment analysis, and the optimal portfolio problem is optimized, and the risk coefficient of different investment ratio is calculated under the condition that the return rate remains unchanged. After comparing with the results of sequential quadratic programming algorithm, it is found that the risk coefficient of the portfolio obtained by genetic algorithm is smaller than that of the portfolio obtained by using the sequential quadratic programming method. It is concluded that the genetic algorithm model is superior to the sequential quadratic programming algorithm in some cases.
【学位授予单位】：兰州大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP18;O221.2;F830.59
本文编号：2369379
[Abstract]:Portfolio optimization problem is a difficult NP problem, and it is difficult to achieve global optimization by common methods. As an efficient parallel global optimization search method, genetic algorithm (GA) has been applied in many fields. In this paper, the genetic algorithm model is introduced into the field of securities investment analysis, and the optimal portfolio problem is optimized, and the risk coefficient of different investment ratio is calculated under the condition that the return rate remains unchanged. After comparing with the results of sequential quadratic programming algorithm, it is found that the risk coefficient of the portfolio obtained by genetic algorithm is smaller than that of the portfolio obtained by using the sequential quadratic programming method. It is concluded that the genetic algorithm model is superior to the sequential quadratic programming algorithm in some cases.
【学位授予单位】：兰州大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP18;O221.2;F830.59
【参考文献】
相关期刊论文 前4条
1 郑美茹;;遗传算法在求解函数优化问题中的应用研究[J];装备制造技术;2011年06期
2 罗勇;陈治亚;;基于改进遗传算法的物流配送路径优化[J];系统工程;2012年08期
3 吴斯;曹炬;;带记忆信息的协同进化算法[J];计算机工程与科学;2008年03期
4 张伟,周群,孙德宝;遗传算法求解最佳证券投资组合[J];数量经济技术经济研究;2001年10期
相关硕士学位论文 前3条
1 林凤伟;遗传算法在网络告警预测中的应用[D];北京邮电大学;2007年
2 魏大松;非线性优化问题的精确罚函数算法研究[D];重庆大学;2007年
3 阮祥;川滇中小地震震源参数的研究[D];中国地震局兰州地震研究所;2007年
，本文编号：2369379
本文链接：https://www.wllwen.com/jingjilunwen/zbyz/2369379.html
