依赖随机参考点的网络均衡配流模型

Network Equilibrium Assignment Model Depended on Stochastic Reference Point

所建立模型明确考虑了随机参考点作为累积前景理论(CPT)描述出行者有限理性路径选择行为的补充,将其定义为随机最短行程时间和可接受系数的乘积.假设出行者遵循路径累积前景最大化原则进行路径选择,建立相应的随机均衡条件及等价的不动点模型.然后,给出基于Probit加载和相继平均法(MSA)的启发式算法,并在小型网络上验证所提出的模型和算法.算例结果表明,依赖随机参考点的交通流模式能够较为真实地再现出行者在路径选择时,同时考虑行程时间均值及随机波动的有限理性行为.对参数进行灵敏度分析,基于CPT得到的路网均衡状态基本上不受行程时间随机波动程度变化的影响,当出行者调整出行时间预算时,均衡状态将随之发生改变.

As a complement for describing travelers＇ boundedly rational route choice behavior by cumulative prospect theory, the stochastic reference point is explicitly considered. It is defined as a product of stochastic shortest travel time and acceptable coefficient. By assuming that travelers follow the rule of maximizing route cumulative prospect to choose the route, corresponding stochastic equilibrium condition and equivalent fixed point model are established. Then, a heuristic algorithm is given based on Probit loading and method of successive averages. The proposed model and algorithm are testified on a small network.Numerical results show that the traffic flow pattern depending on stochastic reference point can virtually represent travelers＇ boundedly rational behavior through average travel time and fluctuation. Due to the sensitivity analysis of parameters, the network equilibrium state from CPT remains unchanged with different travel time fluctuations, but varies when travelers adjust their travel time budgets.