为了刻画出行者的日常择路行为,利用动力学系统方法和不动点理论,建立了一个集成风险规避和认知更新的演化模型,分析了演化过程的稳定性,并在一个简单网络上进行了验证。发现在模拟开始的前15d内,出行时间预算、路径期望出行时间、实际出行时间以及路径流量都出现了较大的波动,但经过大约30d的摸索以后,开始趋向于随机用户均衡状态。分析结果表明:模型所设计的择路演化过程类似于经典的相继平均算法的计算过程,可以确保收敛到稳定状态,并与初始状态参数的取值无关。
In order to depict travelers' day-to-day route choice behaviors, an integrated model considering risk aversion and perception updating was proposed by using dynamics system approach and fixed point theory, its evolution stability was discussed and validated in a simple network. It is pointed that travel time budget, expected route travel time, actual" route travel time and route flow appear a large of turbulence in 15 d at the beginning of simulation, but the traffic pattern evolves to stochastic user equilibrium state after nearly thirty days' travel. Analysis result indicates that the route choice evolution adopted in the model is highly similar to the computational steps of the classical method of successive average(MSA), so the evolution stability can be assured, and the final flow pattern is independent of initial values of all state variables. 2 tabs, 8 figs, 16 refs.