摘要
为有效预测车辆油耗,提高燃油经济性,促进节能减排,提出一种基于XGBoost-MSIWOA-LSTM的车辆油耗优化预测模型。该模型首先采用极端梯度提升树(XGBoost)算法提取车辆油耗特征,以优化模型的输入变量,提高模型的泛化性和鲁棒性。然后,利用多策略改进的鲸鱼优化算法(MSIWOA)对长短期记忆神经网络(LSTM)中的超参数进行自适应寻优,并将优化后的超参数代入LSTM中对车辆油耗进行建模预测。结合实际车辆油耗算例进行对比实验,结果表明,相对于其他对比模型,XGBoost-MSIWOA-LSTM预测模型预测精度更高,对降低车辆油耗具有一定的指导意义。
To effectively predict vehicle fuel consumption,improve fuel economy,promote energy saving and emission reduction,a vehicle fuel consumption optimization prediction model based on XGBoost-MSIWOA-LSTM was proposed.The eXtreme Gradient Boosting tree(XGBoost)algorithm was used to extract vehicle fuel consumption features to optimize the model's input variables and improve the model's generalization ability and robustness.Then,the Multi-Strategy Improved Whale Optimization Algorithm(MSIWOA)was used to adaptively optimize the hyperparameters in the Long Short Term Memory neural network(LSTM),and the optimized hyperparameters were used to model and predict vehicle fuel consumption in the LSTM.Combined with actual vehicle fuel consumption examples for comparative experiments,the results showed that the XGBoost-MSIWOA-LSTM prediction model had higher prediction accuracy than other comparative models,and had certain guiding significance for reducing vehicle fuel consumption.
作者
师国东
胡明茂
宫爱红
龚青山
郭庆贺
谭浩
SHI Guodong;HU Mingmao;GONG Aihong;GONG Qingshan;GUO Qinghe;TAN Hao(S chool of Automotive Intelligent Manufacturing,Hubei University of Automotive Technology,Shiyan 442000,China;Hubei Provincial Key Laboratory of Automotive Power Transmission and Electronic Control,Hubei University of Automotive Technology,Shiyan 442000,China;Dongfeng Commercial Vehicle Co.,Ltd.,Shiyan 442000,China)
出处
《计算机集成制造系统》
北大核心
2025年第9期3467-3484,共18页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(52572402)
湖北省重点研发计划资助项目(2020BAA005)
湖北省教育厅重点资助项目(D20211803)
湖北汽车工业学院博士基金资助项目(BK202001)
湖北省教育厅中青年人才资助项目(Q20221804)
汽车动力传动与电子控制湖北省重点实验室资助项目(ZDK1202205)。
关键词
油耗预测
极端梯度提升树
多策略改进的鲸鱼优化算法
长短期记忆神经网络
自适应寻优
fuel consumption prediction
extreme gradient Boosting tree
multi-strategy improved whale optimization algorithm
long short-term memory neural network
adaptive optimization
