基于免疫记忆克隆算法的指数化投资组合优化构建策略

本文基于克隆选择学说及基于克隆选择学说及生物免疫响应过程的相关机理，提出用于指数化投资的免疫记忆克隆算法，并将其应用于指数化投资组合优化构建模型的求解，旨在探索指数化投资的优化构建策略。文章首先提出多目标的指数化投资组合构建模型。其次，分别设计了适用于指数化投资组合构建策略的抗原、抗体、亲和度函数、克隆选择算子、免疫记忆算子和相应的进化算法。该算法有效避免了传统遗传算法所存在的计算后期解的多样性差、易早熟以及收敛速度慢等缺点。同时，提出了限制投资组合中股票数量的启发式算法。最后，使用包括上证180指数在内的6组世界主要股票市场指数及其成份股的历史数据对模型及算法进行测算，结果表明算法具有良好的求解能力和收敛速度，所建模型的合理性和有效性亦被论证，模型和算法均具有很强的实践价值；

Optimal Portfolio Selection for Index Investing Based on Immune Memory Clonal Algorithm

In order to exploit the optimizing strategy for indexed portfolio selection, an Immune Memory Clonal Algorithm for indexing investment is put forward and applied to optimal indexed portfolio selection based on the clonal selection theory and mechanisms of biological immune response. Indexed portfolio selection with multi-objective is modeled according to the index investing practice. Extra return maximization is included in the model as an objective function. The algorithm design includes antigen, antibody, fitness function, clonal selection operator and immune memory operator. The algorithm effectively overcomes the flaws in the traditional Genetic Algorithm, such as less of result diversity, prematurity and low convergent speed. Meanwhile, a heuristics is designed to limit the number of stocks in the portfolio. Strategy is tested by the historical data of 6 main stock indexes and their component stocks in the world. The results show that: (1)the new evolutionary strategy is capable of improving the search performance significantly both in convergent speed and precision; (2)the indexing portfolio selection model are rational and effective; (3)indexing portfolio selection based on Immune Memory Clonal Algorithm is very helpful in investing practice.