基于累积前景理论的动态交通流演化博弈模型

发布时间：2018-07-18 18:45

【摘要】：交通流演化的内在动力机制是出行者的适应性学习及其诱发的路径选择行为的持续改变,当选择行为趋于稳定时交通系统也将达到或接近均衡状态。首先基于累积前景理论建立了一个用户均衡模型及其等价的变分不等式,在一定约束条件下对模型解的性质进行了讨论;然后将累积前景理论与演化博弈论相结合,利用复制子动态来刻画出行者日常路径选择的博弈学习行为,建立了一个动态交通系统模型,结合稳定性分析证明了当满足一定条件时系统演化能够实现用户均衡;最后通过算例在不同初始状态和不同参照点更新规则下分别展示了交通流的动态演化与用户均衡的实现过程,对相关研究结论进行了验证。本文拓展了传统交通分配模型完全理性假设和均衡分析方法的局限,更加真实、全面的刻画了动态交通系统长期运行特征和规律。
[Abstract]:The intrinsic dynamic mechanism of traffic flow evolution is the continuous change of travelers' adaptive learning and the induced path selection behavior. When the choice behavior tends to stabilize the traffic system will also reach or approach the equilibrium state. Firstly, a user equilibrium model and its equivalent variational inequality are established based on the cumulative foreground theory, and the properties of the solution of the model are discussed under certain constraints, and then the cumulative foreground theory is combined with the evolutionary game theory. By using replicon dynamics to describe the game learning behavior of traveler's daily path selection, a dynamic traffic system model is established, and the stability analysis proves that the system evolution can achieve user equilibrium when certain conditions are satisfied. Finally, the dynamic evolution of traffic flow and the realization of user equilibrium are demonstrated by examples under different initial states and different reference point updating rules, and the relevant conclusions are verified. This paper extends the limitation of the traditional traffic distribution model's complete rationality hypothesis and equilibrium analysis method, and describes the long-term operation characteristics and laws of dynamic traffic system more realistically and comprehensively.
【作者单位】：上海交通大学安泰经济与管理学院;上海交通大学船舶海洋与建筑工程学院;
【基金】：国家自然科学基金资助项目(50978163,71001067)
【分类号】：U491.112

【参考文献】