A Study on the Application of Baysian Approach to the Route Guidance System Planning

本研究的目的在于明确路线引导中驾驶员的路线选择行为对网络流的影响。由路段行程时间估计模型、路线变更模型和路线选择模型组成路线选择行为模型。路段行程时间估计模型利用贝叶斯方法修正了路段行程时间的概率分布。路线变更模型提供了重新考虑路线的机会，路线选择模型决定从现有计划路线变更为新路线。利用开发的路线选择行为模型，对神户市城市场所的路线引导系统进行了仿真。在不实施路由引导固定参数的情况下进行仿真，并将链路交通量与观测值进行比较。结果表明，尽管处理了高峰时间，但链路交通量符合观测值。在实施路线引导时的仿真结果表明，假设的路线引导系统存在交通信息可靠性较低的问题。然而，如果将这些交通信息添加到路径选择行为判断的数据中，诱导车就不会处于不利地位。这一结果表明，当感应车增加时，取消问题的可能性具有启发性。

The purpose of this study is to make the influence which route choice behaivior by driver gives the network flow in route guidance clear. We made route choice behavior model to compose from the link travel time estimate model, the route change model and the route choice model. The link travel time estimate model corrects probability distribution about the travel time using Baysian approach. The route change model gives the opportunity of the route reconsideration and the route choice model decides a change into the new route from the present plan route.Using developed route choice behavior model, it executed the simulation of route guidance system for the city place in Kobe City. It did the simulation when not implementing route guidance to fix a parameter and it compared link traffic volume with the observation value. As a result, in spite of the dealing with the peak time, the link traffic volumes suited observation values. The simulation result when implementing route guidance showed a thing with the low reliability of the traffic information by the assumed route guidance system. However, if making add this traffic information to the data of the judgment for route choice behavior, the induction car never became disadvantageous. This is the result by which the possibility to cancel the problem when an induction car is increased is suggestive.