基于先进交通信息的多目标多类用户降级路网双层优化

Multi-user and multi-objective bi-level optimization of degradable road network based on advanced traffic information

考虑先进的出行者信息系统(ATIS)与可变信息板(VMS)的共同影响,研究混合交通(普通交通与突增的应急交通)在降级路网中的随机均衡分配及路网的性能优化。利用应急交通可靠度和路段负效用2个目标定义广义路径费用,并基于VMS信息效用与路段广义费用设计了新的路径长度(path size,PS)属性公式,进而构造改进的路径长度logit(PSL)模型;建立以降级路段的最差容量利用系数为优化变量的多类用户多目标双层规划模型,其中,下层是基于PSL模型的混合交通随机用户均衡分配模型。利用应急交通的可靠度定义适应度函数,设计以进化算法为主体框架的优化算法,通过求解获得流量分布、ATIS占有率与服从率以及各项性能指标。研究结果表明:模型与算法是可行有效;相对于随机模拟,优化可以大幅度提升应急交通的可靠性,VMS能辅助ATIS以提升应急交通可靠性并降低全路网的总期望出行时间费用;需求增长对降级路网性能会产生较大影响,但VMS仍能在需求增长时有效优化系统性能。

Under the combined effect of advanced traveler information system （ATIS） and variable message sign （VMS）, road network performance optimization and stochastic equilibrium behavior of hybrid traffic （including general traffic and emergency demand increased suddenly） were investigated. Generalized path cost was defined by evacuation reliability and link disutility and a new formula for PS attribute was introduced by incorporating VMS information utility and generalized link cost into the prime formula. Furthermore, an adjusted path size logit （PSL） route choice model was proposed. This led to the establishment of a multi-objective multi-user bi-level programming model with west capacity utilization coefficients of degradable link as optimization variables, in which the lower level was constructed to complete stochastic user equilibrium assignment for hybrid traffic based on the PSL model. For the model, optimization algorithms based on evolutionary algorithm was formulated using evacuation reliability to define fitness function. Flow distribution, ATIS penetration rate and compliance rate, and several road network performance values were obtained by the proposed solution algorithms. Computations and analyses show that the models and algorithms are feasible and effective. The optimization approach can dramatically enhance evacuation reliabilities compared with stochastic simulation. VMS information can assist ATIS to improve evacuation reliability and decrease the total expected travel time. Demand growth has considerable impacts on degradable road network performance; however, VMS can still obviously optimize system performance.