Bi-Objective Optimization Approach for Exclusive Bus Lane Scheduling Design

Abstract

Implementation of an exclusive bus lane (XBL) requires proper planning as it reduces road capacity for nonbus traffic and could possibly worsen traffic congestion in urban cities. Exclusive bus lane scheduling (XBLS) has been implemented in several cities but a systematic methodology on the design of such schedules is yet to be made available. This paper proposes a biobjective optimization model to optimize a bus lane schedule taking into consideration bus and nonbus traffic performances. A hybrid nondominated sorting genetic algorithm (NSGA II) with a microscopic traffic simulation model was adopted to solve the optimization model. Through the analyses presented in this paper, it was found that the proposed model is capable of producing XBL schedules to relieve traffic congestion, while being sensitive to key parameters (such as population size, travel demand level, and minimum duration for implementation). Pareto solutions from the model allow engineers to decide the appropriate trade-off for improving bus performance without overdeteriorating nonbus traffic performance. This is crucial for sustainable transport planning in which the performances of multimode transport systems need to be balanced.