基于改进遗传BP算法的产业产值中单位GDP碳排放量的预测研究

Prediction of carbon emissions from energy consumption based on the improved genetic algorithm BP's industrial output value

针对产业产值中单位GDP碳排放量的预测问题,系统分析了产业产值与能源消耗碳排放量的关系,在传统遗传算法的基础上,根据动态神经网络模型优化算法过程,克服了传统遗传算法的无动态性,提出了改进遗传BP算法.根据2005—2014年北京等10个省市第三产业的单位GDP碳排放量相关数据资料,利用改进遗传BP算法预测各省市在2015年的单位GDP碳排放量,其预测值与真实值的平均误差值为-0.03,其平均相对误差仅为1.06%,说明该算法在数值预测方面的有效性.

For the prediction problem of the carbon emissions of per unit GDP in connection with industrial output value, this paper analyzes the relationship between the industrial output and the carbon emissions from energy con- sumption. On the basis of the traditional genetic algorithm, it optimizes the algorithmic process in line with the dy- namic neural network model, overcomes the defect of the traditional genetic algorithm, and proposes an improved genetic algorithm BP. Finally, according to the carbon emissions of per unit GDP in the tertiary industry in 2010 - 2014 in Beijing and other ten provinces, it uses the improved genetic algorithm BP to predict carbon emissions of per unit GDP in 2015, its predicted value and the actual value of the average error value is -0.03, while the aver- age relative error is only 1.06%, indicating that the algorithm is effective in terms of numerical prediction.