Bilevel Programming Model for Locating Park-and-Ride Facilities

Abstract

This paper develops a bilevel programming model to locate park-and-ride (P&R) facilities with the capability of capturing the interactions between decision makers and commuters. The upper-level model represents the authorities’ locating decision with the objective of maximizing the total social welfare (SW); the lower-level model is a stochastic user equilibrium one simulating commuters’ reactive behaviors, which in turn affects the P&R location decision at the upper level. The paper designs a scenario to compare the solutions under the optimization criterion of SW with that under two others: consumer surplus and producer surplus. Results indicate that adopting SW maximization as the optimization objective leads to a win-win outcome for both P&R service providers and users. The paper uses Chengdu City (in China) as a case study, and it is demonstrated that the proposed P&R locating framework is applicable to a large-scale network and compatible with the regional transportation planning procedure.