Optimization of user and operator cost for large-scale transit network

A methodology for optimizing transit networks based on both passenger and operator costs is presented. Given information on transit demand, street network, and a set of feasibility constraints of a transit service area, the methodology searches for the transit network that best fits design goals through the minimization of a total cost objective function. The goal is to provide an effective mathematical solution procedure with minimal reliance on heuristics to solve large-scale transit network optimization problems. The methodology consists of a representation of transient route network and headway search spaces; a normalized, dimensionless total cost function; and a stochastic global search scheme that combines simulated annealing, tabu, greedy, and bisection search methods. The methodology has been tested with published benchmark problems and applied to a large-scale realistic network optimization problem. The results show that the methodology is capable of producing improved solutions to large-scale transit network design problems with reasonable computing resources.