Optimizing route choice in multimodal transportation networks
Advanced traveler information systems provide travelers with pre-trip and en route travel information necessary to improve the trip decision making process based on various criteria (e.g., avoiding the negative impacts of traffic congestion, selecting specific travel modes, etc.). This study investigates an adaptive routing methodology for multimodal transportation networks. To integrate transit networks, the model takes into account both the predefined timetables of public transportation services and the variability of travel times. A graph theory based methodology is proposed to capture travel behavior within a multimodal network. The study advances a routing algorithm based on Markov decision processes. Special network modeling elements were defined to allow the developed algorithm to select the most efficient transportation mode at each junction along a given route. The proposed methodology is applied to a small real-world network located in the central business district area of Montreal, Quebec. The network includes bus, subway, and bicycle transportation facilities. The simulations were run under the assumption that users do not use private vehicles to travel between arbitrary selected origin and destination points. The developed routing algorithm was applied to several simulation scenarios. The results identified what is the most efficient combination of transportation modes that the travelers have to use given certain traffic and transit service conditions. Larger and more complex networks of motorized and non-motorized modes with stochastic properties will be investigated in subsequent work.