Traffic Assignment Model And System Optimization On A Multimodal Transportation Network Including Carpooling

As people travel more frequently and the increasing urban cars, the urban traffic congestion is becoming more and more serious. Carpooling is considered to be an effective way to ease traffic congestion because it can share the capacity of the cars and road resources. With the development and application of intelligent transportation system, carpooling has gradually become a new type of transportation mode which was accepted and adopted by people. Carpooling will certainly affect the traffic equilibrium. In order to encourage people to choose carpooling, many cities provide HOV/HOT lanes for traveler, aiming to guide people’s traffic patterns and travel path. Traffic planners need to systematically think how to design the HOV/HOT lanes.On the basis of generalizing traffic equilibrium models, assume that people choose private transportation or public transportation first, then model, finally choose the path, and according to the equilibrium condition find the equivalent variational inequality model. Design a numerical example including three modes:carpooling, solo, bus. Through the example analysis find carpooling can greatly improve the performance of urban traffic network performance, reducing the total travel time of the system. With the rising fuel price, the ration of choosing carpooling increased at first and then decreased. HOV lanes and HOT lanes will affect the flow of the traffic patterns share, they can optimize the performance of the system and can also make the whole system performance worse. Transportation planners need systematic thinking to make decisions.A bi-level model is designed to minimize total system travel cost and to find the optimal design of HOV/HOT lanes. GA is used to solve this problem. In a multimodal transportation network including carpooling, solo and bus, find out the optimal design of HOV lanes to minimize the total travel cost in the system. In multimodal transportation network, there may be a peak time or trough time of traffic demand. The utilization of HOV lanes in trough time which was designed for peak time may not be high, so we will change these HOV lanes into HOT lanes in trough time. Through the model and algorithm, we find the optimal HOT lanes charging scheme to minimize the total travel cost.