Abstract
The research presented in this paper proposes a Particle Swarm Optimization (PSO) approach for solving the transit network design problem in large urban areas. The solving procedure is divided in two main phases: in the first step, a heuristic route generation algorithm provides a preliminary set of feasible and comparable routes, according to three different design criteria; in the second step, the optimal network configuration is found by applying a PSO-based procedure. This study presents a comparison between the results of the PSO approach and the results of a procedure based on Genetic Algorithms (GAs). Both methods were tested on a real-size network in Rome, in order to compare their efficiency and effectiveness in optimal transit network calculation. The results show that the PSO approach promises more efficiency and effectiveness than GAs in producing optimal solutions.
Highlights
In recent decades, many policies for urban transport planning have been proposed; these policies aim at mitigating the negative effects of the overuse of cars in urban centres with regard to air and noise pollution, energy consumption, safety, and traffic congestion.An effective strategy to achieve sustainability is to encourage mode switching from car to public transport by increasing its accessibility and reliability
The research presented in this paper proposes a Particle Swarm Optimization (PSO) approach for solving the transit network design problem in large urban areas
This paper proposes a PSO approach procedure for solving the bus network design problem
Summary
Many policies for urban transport planning have been proposed; these policies aim at mitigating the negative effects of the overuse of cars in urban centres with regard to air and noise pollution, energy consumption, safety, and traffic congestion. A set of potential stops is created using a clustering criterion; routes are generated by applying a modified shortest-path procedure based on Newton’s gravity theory This set of candidate routes (organized into a hierarchy of mass, feeder, and local routes) is the main input for the last step of the process: a metaheuristic technique, deriving from an ACO algorithm hybridized by a GA, finds the optimal network; both the budget constraints and level of service standards are fulfilled. Their method has been tested on the Winnipeg network and applied on the Chicago extra-urban rail network (almost 13,000 nodes, 52,000 links, 650 stops, and 500 routes).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
