Real-Time Scheduling for a Variable-Route Bus System

Abstract

Traditional buses operate on an outdated fixed-route system that is often characterized by inefficiencies such as trips to empty bus stops and long waiting time at the station. However, the prevalence of personal mobile devices provides more opportunities than ever for passengers to interact with the buses, giving rise to the possibility of personalized travel routines. In this thesis, we propose a variable-route system that can customize and optimize bus routes in real time based on the passenger requests and their characteristics. Using a computer simulated example, we demonstrate that the variable routing system reduce the total waiting time on bus stop and the total travel time on the buses by a significant margin. Furthermore, we show that by leveraging genetic algorithm, we can reduce the computation time by as much as 90\% while still maintaining a significant reduction in waiting time and traveling time. This thesis provides an initial performance check for the optimal bus route scheduler and highlights an implementation that can achieve significant time reduction with much less computation time.