基于遗传算法的指数复制方法及其实证研究

An empirical study of index replication method based on genetic algorithm

指数化投资策略是证券市场主要投资方法和投资策略之一,其核心内容就是构建一个能够完全复制指数走势的跟踪组合,所以研究指数复制方法具有重要的理论意义和巨大的应用价值.对标准的遗传算法进行改造,重新设计了遗传算法的编码方式,适应度函数以及遗传算子,采用最优保存策略加快算法的收敛效率.使用MATLAB语言实现算法,以沪深300作为目标指数,对历史数据进行回测,从实证结果来看,所设计的算法取得不错的跟踪效果.通过对遗传算法和传统的市值抽样法和行业分层抽样法进行对比,发现遗传算法的跟踪效果略优于传统的指数复制方法,遗传算法在指数复制领域有广阔的应用前景.

Indexation investment strategy is one of the main investment strategies in the securities market. The main goal of index investment is to build a stock portfolio to replicate the standard genetic algorithm,redesignsed its coding,fitness function and genetic operator. In the meanwhile,we used elitist strategy to accelerate the convergence of the algorithm. We used MATLAB language to implementar algorithm and applied to SH300 index back testing with historical data. From the empirical results,our redesigned algorithm achieved good tracking effect. We compared our algorithm with traditional market capitalization sampling and industry sampling stratified sampling method,and found that the track performance of genetic algorithm is better than the conventional index replication methods, our study shows that the genetic algorithm in the index replication has broad applications field.