大型交通网络OD矩阵推算方法研究

Research on Method of Estimating O D Matrices for Large Sized Transportation Network

ＯＤ矩阵是进行交通规划及管理的基础数据．本文首先研究了推算ＯＤ矩阵所需基本信息的获取方法．详细评述了现有ＯＤ矩阵推算模型，针对极大熵模型对路段交通量完全置信的不足，提出了极大熵修正模型，使模型能考虑交通调查产生的误差及交通分配引起的误差，极大提高了推算结果的精度．此外，本文把推算模型求解的立足点从路段转移到路径，提出了适用于大型网络ＯＤ矩阵推算的路径迭代算法．最后，本文采用山东省公路网实测ＯＤ矩阵及路段流量数据，对推算模型进行了检验，并证实了推算模型的可靠性及可行性．

Origin destination matrices are the basic data in transportation planning and management. Firstly, the method obtaining basic information which is required for estimating O D matrices is discussed. Secondly, current O D estumation models are analyzed in detail, and an improved entroppy maximising (EM) model is presented in view of shortcoming of existing EM model. The improved model can consider link volume errors which are caused by traffic count or traffic assignment, and the precision of estimated O D matrices can be improved largely. Thirdly, a Path iteration Algorithm of solving EM model on large sized transportation network is presented according to transferring analysis element from links to paths. Finally, the estimation model is examined on the basis of real O D matrices and link volumes of Shan Dong Provincial highway metwork, and feasibility and reliability of estimation model is confirmed.