The invention provides a short-time road network MFD parameter prediction method based on LSTM, and belongs to the field of traffic state prediction. Comprising the following steps: respectively creating respective LSTM models by taking two parameters of road network weighted flow qw and road network weighted density kw in short-time road network MFD parameters as prediction targets so as to determine a loss function; initializing parameters in the loss function, and continuously updating the parameters in the loss function until the loss function is converged to complete model training; the method comprises the following steps: establishing a multi-objective optimization model by using a multi-objective optimization algorithm GAMOPSO, importing data trained by the model into the multi-objective optimization model for prediction to obtain two groups of optimal LSTM structures and hyper-parameters, and applying the two groups of optimal LSTM structures and hyper-parameters to obtain two parameter combinations of optimal road network weighted flow qw and road network weighted density kw, namely a final prediction result. According to the invention, the parameters of the short-time road network MFD can be accurately predicted, the defect of over-metering is effectively avoided, the potential traffic state can be found in time, and the travel efficiency is improved.

    本发明提供了一种基于LSTM的短时路网MFD参数预测方法,属于交通状态预测领域。包括:以短时路网MFD参数中的路网加权流量qw、路网加权密度kw两个参数作为预测目标分别创建各自的LSTM模型,以确定损失函数;初始化损失函数中的参数,并对损失函数中的参数进行不断更新,直至损失函数收敛完成模型训练;利用多目标优化算法GAMOPSO建立多目标优化模型,将模型训练的数据导入多目标优化模型中进行预测,获得最佳的两组LSTM结构及超参数,并将其应用,以获得最佳路网加权流量qw、路网加权密度kw两个参数组合,即为最终预测的结果。通过本发明能够精准预测短时路网MFD的参数,有效避免过于集计性的缺陷,以及时发现潜在的交通状态,提高出行效率。


    Access

    Download


    Export, share and cite



    Title :

    Short-time road network MFD parameter prediction method based on LSTM


    Additional title:

    基于LSTM的短时路网MFD参数预测方法


    Contributors:
    LIN XUANHUA (author) / TAN CHAOJIAN (author) / LIN XIAOHUI (author) / HUANG LIANG (author) / LONG QINGWEN (author) / HUANG MINGJING (author) / LIN PEIHAO (author) / HAYASHI MASASHI (author) / YE JINKUI (author)

    Publication date :

    2024-09-13


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung




    D-LSTM: Short-Term Road Traffic Speed Prediction Model Based on GPS Positioning Data

    Meng, Xianwei / Fu, Hao / Peng, Liqun et al. | IEEE | 2022


    Short-Term Traffic Flow Prediction Based on Graph Convolutional Network Embedded LSTM

    Huang, Yanguo / Zhang, Shuo / Wen, Junlin et al. | TIBKAT | 2020


    Aeroengine Gas Path Parameter Trend Prediction Based on LSTM

    Zhou, Qifeng / Chen, Xiaonan / Qing, Xinlin et al. | SAE Technical Papers | 2023


    Short traffic flow prediction method based on ARIMA and LSTM hybrid neural network

    WANG WEI / ZHOU WEI / HUA XUEDONG et al. | European Patent Office | 2021

    Free access