一种改进的GEP方法及其在演化建模预测中的应用

Application of a novel GEP algorithm in evolutionary modeling and forecasting

为了提高预测的准确性,结合基于突变算子的基因表达式和一种基于群体搜索技术的演化算法,提出了改进的GEP方法,并用其对数据进行演化建模。实例测试的结果表明,使用改进的GEP方法得到的模型要优于GP和单纯的GEP方法得到的模型,本方法具有良好的拟合度和预测精度。

To improve the veracity of forecasting, a hybrid GEP algorithm was presented, which combined GEP based on mutation operator and evolutionary algorithm based on colony-search strategy. The new GEP algorithm was applied to establish models with testing data. The applications show that the models set up by the new GEP algorithm are superior to the models set up by GP and simple GEP. The new algorithm has high-blooded forecasting accuracy.