Optimal Toll-Area and Toll-Level Design of Road Networks on Baseof Second-Best Multi-Class Based Congestion Pricing Theory

Congestion pricing strategy as a traffic demand management measure is widely accepted by the public. However, there is little focus on the research of optimal toll-areas. This paper puts forward the Second-best multi-class based bi-level planning model of road network congestion pricing in elastic demand, where the upper level model attempts to maximize the network social welfare by trying different tolls on some roads and the lower level model describes the route choice behavior of network users. A Multi-class user equilibrium Frank-Wolfe algorithm is used for the lower level model and a differential quotient algorithm of bi-level planning model is employed to find the optimal solution after function evaluations. Finally, on basis of the above model and algorithm, the optimal toll area is computed its toll level is simultaneously designed.