Simultaneous Optimization of Passive Transit Priority Signals and Lane Allocation

Abstract

This paper presents an integrated optimization model for the concurrent design of passive bus priority signal settings and lane allocation including lane markings for general traffic and exclusive bus lanes at isolated intersections. Traditional passive transit signal priority (TSP) strategies considered both lane markings for general traffic and exclusive bus lane settings as exogenous inputs. However, in our method, two modes of transportation (passenger cars and buses) have been considered in a centralized framework while exclusive bus lane settings, lane markings, and the optimal passive transit priority signal settings can be optimized concurrently. Minimization of the weighted delay of buses and cars is adopted as the aim of the integrated optimization model. Based on traditional National Electrical Manufacturers Association (NEMA) phase constructure, a precedence graph was employed, by which the mathematical optimization model is formulated. A set of restrictions have been set up to verify efficiency and safety for the resulting optimal lane markings and signal settings. A simulated annealing (SA) algorithm is designed to resolve the mix-integer-nonlinear-programming (MINLP) problem. In order to demonstrate the efficacy of the method proposed, numerical examples have been provided. The results from sensitivity analysis show that the proposed method can generate sub-optimal solutions for the intersection of different combinations of car demand, bus demand, and left turn ratios.