||In this thesis, the problems of pricing are investigated on networks with heterogeneous users. I seek the possibility and conditions through an anonymous toll scheme to achieve a simultaneously improvement of multi-criteria system performance. First, I look into the possibility and conditions of a Pareto-improving and Revenue-neutral congestion pricing scheme on a simple bi-mode network: this scheme aims at improving system performance, making every individual user better off (includes increasing social welfare) and having zero total revenue simultaneously. Second, comparing with the “do-nothing” case, I look into the possibility of Pareto-improving of system performances measured in both time and monetary units (the system time and the system cost) through a congestion pricing scheme in general networks. In the first part, since Pareto-improving and Revenue-neutral scheme suggested here charges transit users negative tolls and automobile users positive tolls, it can be a proper way to implement congestion pricing and transit subsidy in one step; also it will possibly improve system performance, increase social welfare and offset the inequity for the poor. I concern with two problems. First, whether or not Pareto-improving and Revenue-neutral pricing scheme exists for certain targeted modal split flow pattern? Second, what are the different Pareto-improving situations when a bi-mode transportation system serves for travel groups with different VOT distributions respectively? For a general VOT distribution, the condition that everyone will be better off is found. Furthermore, the impacts of details of different travel communities’ VOT distributions on the better off situations after Revenue-neutral toll are studied and the better off situations can be classified into three cases. Then I focused on travel groups with uniform VOT distributions. A continuous range in terms of the targeted highway flow always exists to realize everyone better off through Revenue-neutral pricing. The larger the VOT range is, the smaller the range is. A common range that the system time is decreased is exactly the range that Pareto-improving and Revenue-neutral pricing scheme exists, independent of VOT distribution. However, considering the increasing and decreasing property of the loss of the most adverse user among the community, a large VOT range does not always mean the adverse inequitable situation after Revenue-neutral pricing. When the reduction of travel time on highway through pricing is large enough, the large widening rich-poor gap will give more benefit to the most adverse user. In the second part, comparing with the untolled user equilibrium situation, the necessary condition of Pareto-improving of the total system travel cost and the total system travel time through a road pricing scheme in general networks is found. I give a classification of the untolled user equilibrium situations which are collectively exhaustive and mutually exclusive. Two among the classification are identified as extreme situations where the two objectives might completely conflict with each other and show that, for general cases except the two extreme situations, the system time and cost can be reduced simultaneously as long as the system time can be reduced separately.