基于BPNN-MOP模型的区域公路网合理规模预测研究

区域公路网规模是保障交通供给能力的基础，确定合理的路网规模是调节交通供需平衡的关键. 本文从路网发展数量和质量两方面，里程、密度、等级结构三项指标出发，建立了基于BP神经网络与马尔可夫链的公路网里程规模组合预测模型，并采用多目标规划方法对公路网等级结构进行了优化研究. 该方法能够充分挖掘公路网规模演变的宏观调整与微观波动规律，提高了预测的精度和可靠性. 最后本文结合实际案例，对区域特征年公路网的合理规模及结构进行了分析预测. 结果表明，组合预测方法能够较科学、客观地反映公路网发展的数量特征和质量要求，具有一定的理论价值和现实意义.

Reasonable Scale Foresting of Regional Highway Network Based on BPNN-MOP

The scale of the regional highway network is the basis of traffic supply determination, and whether it is reasonable or not has strong relaptionships with the balance between traffic supply and demand. Considering the length, density, and hierarchical structure of highway network, the paper first develops the predicted model based on the BP neural networks and Markov chains. Then, a multi-objective programming model is used to define the optimum structural of highway. The method can fully describe the macro and micro fluctuation laws of highway network scale evolution, and improve the prediction accuracy and reliability as well. Finally, this paper estimates the highway scale and reasonable hierarchical structure in case of one province for the year of 2010, 2015, and 2020. The results prove that the model has high accuracy and reasonableness and have some theoretical and practical significance.