最大熵模型在公共交通分布预测中的应用

探讨了在无现状起讫点矩阵(O-D矩阵)的情况下如何应用最大熵模型较好地描述公交出行分布。利用公交定位技术和收费系统的有效信息,以乘客公交出行距离分布为约束条件,将对应状态数最多的出行分布视为预测的出行分布,以乘客的微观行为来反映公交分布的宏观状态。提出改进后熵模型的参数标定方法;并通过实例分析与佐佐木模型的预测结果进行比较,结果表明改进后的最大熵模型适用性较强,在公共交通分布预测中有很好的应用前景。

Application of Entropy-maximizing (EM) Modelin Public Traffic Distribution Forecast

The paper researched the better way to describe the traffic distribution of public transport by applying entropy-maximizing (EM) model without the present O-D Matrix. Making full use of the effective information of global position system and bus charge system, the improved model added to the passenger trip length distribution as a constraint condition and regarded traffic distribution with the most number of corresponding states as forecasting distribution, thus well reflecting the macro state of public transport distribution by the micro behavior of passengers. The parameters calibration method was proposed to actual application. Through analyzing and calculating an example, the forecasting results were compared with those of sasaki model. The results show that the application of improved EM model will have a promising prospect for the public transportation distribution forecast.