基于多Agent的早高峰公交通勤者仿真研究

发布时间：2018-04-04 14:04

本文选题：公共交通　切入点：出行均衡　出处：《天津大学》2014年硕士论文

【摘要】：城市居民出行的聚集现象越加严重,尤其是在早晚高峰时段,由此而导致的城市交通拥挤问题已严重影响了居民的生活并且带来了空气污染及停车用地匮乏等问题。对通勤者出发时间的研究作为交通需求管理中的重要组成部分,能够帮助我们充分的理解出行者出发时间选择的影响因素和行为机制,是设计和评估交通政策的先决条件。结合多Agent仿真技术及强化学习Bush Mosteller算法,建立城市早高峰公交通勤出发时间选择仿真模型。通过对比相同条件下经典解析模型的结果验证了模型的正确性,并且再现了均衡的形成过程。从通勤者自身因素和环境政策因素方面出发进行了两类仿真实验,在关于通勤者自身因素的实验中,主要考虑了通勤者的异质性及有限记忆的特性;而在关于环境政策因素的实验中,重点考察了公交票价“峰前免费”政策及“公交优先”政策对出发时间均衡的影响。本研究加深了对通勤者出发时间选择行为和早高峰通勤均衡现象的理解,使用的方法也为探索复杂交通现象的形成和演化过程提供了一种有效的途径。
[Abstract]:The phenomenon of urban residents' travel agglomeration is more and more serious, especially in the morning and evening rush hour, resulting in urban traffic congestion has seriously affected the lives of residents and brought air pollution and lack of parking space and other problems.As an important part of traffic demand management, the study of commuter departure time can help us fully understand the influencing factors and behavior mechanism of traveler departure time choice, which is a prerequisite for the design and evaluation of transportation policy.Combined with multiple Agent simulation technology and reinforcement learning Bush Mosteller algorithm, a simulation model for selecting the departure time of urban morning rush bus commute is established.The correctness of the model is verified by comparing the results of classical analytical models under the same conditions, and the formation process of equilibrium is reproduced.Two kinds of simulation experiments are carried out from the aspects of commuter's own factors and environmental policy factors. In the experiment of commuter's own factors, the heterogeneity of commuters and the characteristics of limited memory are considered.In the experiment of environmental policy factors, the influence of bus fare "free before peak" policy and "bus priority" policy on departure time balance is investigated.This study has deepened the understanding of commuters' departure time choice behavior and the phenomenon of commuting equilibrium in early rush hours. The methods used in this study also provide an effective way to explore the formation and evolution of complex traffic phenomena.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.17

【参考文献】