A Multi-objective Optimization of Signal Timings for Two-Stage Midblock Pedestrian Crosswalk

摘要

With the goal to improve convenience and safety for pedestrians, signalized two-stagemidblock pedestrian crosswalks (MPC) have been increasingly installed in highlypopulated areas in developing countries such as China. This paper presents amulti-objective optimization model and its solution algorithm for optimal control of atwo-stage midblock crosswalk which consists of different traffic modes (e.g. walking anddriving). With a central refuge island as the safe area for pedestrians to stop, the proposedmodel aims to produce the optimal signal timings at the crosswalk to best accommodateboth vehicular traffic and pedestrian movements, and to concurrently adjust offsets ofpedestrian signals of the two stages to minimize pedestrian delays and relieve thecongestion at the central refuge island. Different from previous studies, the proposedmodel and algorithm have the following three critical features: (1) modeling thepedestrian delays explicitly at the two-stage controlled crosswalk, including delays atboth the curb side and the central refuge island; (2) modeling the number of waitingpedestrians on central refuge island with respect to the change of signal timings; and (3)developing a multi-objective optimization approach to allow traffic engineers toefficiently explore the control effectiveness of mid-block crosswalk under differentpriorities between the vehicular and pedestrian traffic. A heuristic based on NSGA-II wasdesigned to solve the model and generate the Pareto solution set for signal timings. Casestudy results have shown the promising property of the proposed model to assist trafficpractitioners, researchers, and authorities in properly designing signal timing plans andcentral refuge island for two-stage midblock pedestrian crosswalks.