基于参数二次规划和逆向优化的超越股票市场指数的实证研究

Directly Evaluating Genetic Algorithms for Portfolio Optimization in the Context of Big Data

习近平总书记在2015年底提到股票市场的发展,对股票市场的发展提出了新的要求,股票市场的健康发展需要投资者主动控制投资组合的风险,这促使投资组合在我国股票市场的应用更加广泛,更对投资组合模型的效果研究提出了新的要求。为了科学系统地探索投资组合模型在我国股票市场中的应用效果,本文选取了自2013年1月至2015年12月中由证监会划分的18个行业中流通市值最大的股票,以此建立带上界约束条件的投资组合模型。针对股票市场指数的收益率,本文创新性地使用参数二次规划和逆向优化计算出精确的权重向量的函数表达式,求解出特定收益率下最优投资组合的权重向量,进而对比在下一时间阶段内我们所构造的投资组合与股票市场指数收益率。12个实验组的分析结果表明,基于参数二次规划和逆向优化下的投资组合收益率基本超越了同时期的股票市场指数的收益率,实证表明投资组合模型在我国股票市场有着一定的应用价值。

President Xi Jinping put forward new requirement of development of stock market when he mentioned stock market development on the end of 2015. Investors control investment portfolio actively is important for developing stock market healthily,which makes investment portfolio to be widely used in stock market. At the same time,the significance of investment portfolio makes the research on portfolio model to achieve a higher level. To explore the effectiveness of applying portfolio model to Chinese stock market,this study selects the largest stock from 18 industries divided by CSRC from January 2013 to December 2015, combining with portfolio model with upper bound constraint. The innovation of this study is to use parametric quadratic programming and inverse optimization to calculate the exact weight vector function in the optimal portfolio according to a specific expected return,and we compare the return of the portfolio and the stock market index. The 12 control experimental groups' result shows that based on parametric quadratic programming and inverse portfolio optimization, the return of portfolio exceeds the return of stock market index. This study proves that the portfolio model has a certain application value in Chinese stock market.