RDWKCPSO-PCA-BPNN的汽车燃油消耗预测  

Automobile fuel consumption prediction based on RDWKCPSO-PCA-BPNN

作　　者：姜平[1] 祖春胜 李晓勇[1] 

机构地区：[1]合肥工业大学机械与汽车工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2016年第1期7-13,共7页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(51178158);安徽省自然科学基金资助项目(1508085ME94)

摘　　要：针对不同乘用车综合工况下理论百公里燃油消耗数据,文章提出了一种基于主成分分析(principal component analysis,PCA)和BP神经网络(back propagation neural network,BPNN)的燃油消耗预测模型;通过PCA方法对选取影响理论燃油消耗的24个因素进行压缩,简化模型结构,再利用BPNN算法,构建燃油消耗预测模型;由于神经网络中的权值和阈值对预测模型效果影响较大,采用基于随机动态惯性权重和Kent映射的混沌粒子群算法(RDWKCPSO)优化PCA-BPNN模型中的权值和阈值。对3种标准函数的寻优测试结果表明,RDWKCPSO优化算法更容易跳出局部最优实现全局寻优,提高了模型适应能力及预测精度。On the basis of the theoretical one-hundred-kilometer fuel consumption data under comprehensive conditions of different passenger car, the fuel consumption prediction model based on principal component analysis（PCA） and back propagation neural network（BPNN） is proposed. In order to simplify the model structure, the 24 factors that affect the theoretical fuel consumption are compressed by PCA, then the fuel consumption prediction model is established by using BPNN algorithm. Because the weight and threshold of neural network have a greater impact on the model prediction effect, the weight and threshold of PCA-BPNN model is optimized by applying the chaotic particle swarm optimization algorithm based on random dynamic inertia weight and Kent map（RDWKCPSO）. The optimization test results of three kinds of standard function show that the RDWKCPSO optimization algorithm is more likely to jump out of local optimization to find the global optimization and the model adaptability and prediction precision are improved.

关 键 词：BP神经网络 权值和阈值 混沌粒子群算法 主成分分析 燃油消耗预测 Kent映射 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] U462.34[自动化与计算机技术—计算机科学与技术]