Modeling and managing the morning commute problem with park-and-ride-sharing

Abstract

In this paper, park-and-ride-sharing (P&RS) is proposed to manage morning peak hour congestion in a monocentric linear city, where there is a multilane highway with a single bottleneck connecting the residential area and the central business district (CBD). To realize ride-sharing matching, a ride-sharing parking lot (RSPL) and a ride-sharing meeting point (RSMP) are set up somewhere between the origin and the bottleneck. To encourage ride-sharing participation, the lanes of the highway are physically divided into high occupancy vehicle (HOV) and general purpose (GP) lanes for only ride-sharing vehicles and non-ride-sharing vehicles, respectively. The user equilibrium principle in terms of generalized travel cost is adopted to formulate commuters’ mode choice and arrival time choice at the RSMP. It is demonstrated that there exist three possible equilibrium ride-sharing matching patterns and three possible equilibrium traffic flow patterns. All possible equilibrium ride-sharing matching and traffic flow patterns are analytically derived. In addition, to minimize total system cost (TSC), three break-even parking charge schemes are designed under the case that all parking lots are managed by the government. The optimal allocation scheme of bottleneck capacity to HOV lanes under each parking charge scheme is analytically derived. Finally, numerical examples are provided to illustrate the properties of the proposed bottleneck model with P&RS and the effectiveness of the designed parking charge schemes.