Date of Award

December 2019

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Jie Yu

Committee Members

Yue Liu, Matthew Petering, Chao Zhu, Ying Wang


Traffic bottleneck is defined as a disruption of traffic flow through a freeway or an arterial, which can be divided as two categories: stationary bottleneck and moving bottleneck. The stationary bottleneck is mainly formed by the lane drops in the multi-lane roadways, while the moving bottleneck are due to the very slowing moving vehicles which disrupt the traffic flow. Traffic bottlenecks not only impact the mobility, but also potentially cause safety issues.

Traditional strategies for eliminating bottlenecks mainly focus on expanding supply including road widening, green interval lengthening and optimization of intersection channelization. In addition, a few macroscopic methods are also made to optimize the traffic demand such as routing optimization, but these studies have some drawbacks due to the limitations of times and methodologies.

Therefore, this research utilizes the Connected and Autonomous Vehicles (CAV) technology to develop several cooperative trajectory optimization models for mitigating mobility and safety impact caused by the urban bottlenecks. The multi-phases algorithms is developed to help solve the model, where a multi-stage-based nonlinear programming procedure is developed in the first phase to search trajectories that eliminate the conflicts in the bottleneck and minimize the travel time and the remaining ones refine the trajectories with a mixed integer linear programming to minimize idling time of vehicles, so that fuel consumption and emissions can be lowered down. Sensitivity analyses are also conducted towards those models and they imply that several indices may significantly impact the effectiveness and even cause the models lose efficacy under extreme values.

Various illustrative examples and sensitivity analyses are provided to validate the proposed models. Results indicate that (a) the model is effective to mitigate the mobility and safety impact of bottleneck under the appropriate environment; (b) the model could simultaneously optimize the trajectories of vehicles to lower down fuel consumption and emissions; (c) Some environment indices may significantly impact the models, and even cause the model to lose efficacy under extreme values. Application of the developed models under a real-world case illustrates its capability of providing informative quantitative measures to support decisions in designing, maintaining, and operating the intelligent transportation management.