体液免疫算法及其对证券组合投资分析的应用  

Humoral Immune Algorithm and Its Application to Portfolio Selection

作　　者：张著洪[1] 胡支军[1] 

机构地区：[1]贵州大学理学院数学系,贵阳550025

出　　处：《系统工程理论与实践》2006年第11期84-91,共8页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(60565002);贵州省优秀人才省长基金(20040706);贵州省自然科学基金(20052002);贵州大学博士启动基金(2004001)

摘　　要：基于Jia&Dyer的一般性失望模型,给出一种新的非对称风险度量方法,建立该风险度量下考虑证券最小交易单位约束的组合投资二次整数规划模型;进而依据体液免疫原理设计实用、简单的新体液免疫算法,并寻求该模型的最优方案.算法设计中引入优秀抗体演化操作,搜集和更新进化中最好解,以及建立能增强群体多样性及具有较强整体、局部、并行搜索能力的免疫操作,从多方位搜索最优解.实证及比较表明,所获算法的整体和局部搜索能力强、能快速获取最优投资决策方案,所建模型的合理性和有效性被论证.A portfolio selection quadratic integer programming model with minimum transaction lots is developed through designing a new asymmetric risk measurement based on the generalized disappointment model proposed by Jia and Dyer. Further, associated with humoral immune theory, a novel and available humoral immune algorithm is proposed to solve the model. In the design of the algorithm, the key is to introduce an excellent antibody sot - updating operator to collect and update the good solutions found from the search process, while establishing several kinds of .immune operations being capable of improving diversity of population and strengthening the capability of global, local, and parallel search so that the algorithm can search the optimal solution from different directions. Practical application and numerically comparative resuks, illustrating the reasonability and availability of the model, show that the proposed algorithm with strongly global and local search capability can obtain rapidly the optimal investment decision-making scheme of the model.

关 键 词：风险度量 投资组合 免疫算法 体液免疫 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]