Resumen
Many studies have been done to model and design park and ride scheme, since park and ride has long been considered as a strategy to alleviate traffic congestion and reduce emission. However, due to the difficulty in modelling an urban transport network system, research on park and ride still needs to overcome some limitations. This paper proposes a robust model by combining the driver's mode choice and route choice, which are based on a combined cross nest logit (CNL) and user equilibrium (UE) model. Mathematical programming and varitional inequality are used to solve a network optimization. The main contributions of this paper are: (1)considering travel time uncertainty on roads which can affect both modal split and route choice. (2)taking multi-class demands into consideration because different people have different requirements for the travel time and level of service. In other words, mean-excess travel time (METT) model are proposed in this paper to depict uncertainty environment. Heterogeneity is considered by commuters' variations of METT parameter which depends on travelers' trip characteristic and income level.