System optimal routing of traffic flows with user constraints using linear programming

For static traffic assignment problems, it is well known that (1) for some users the experienced travel time in a system optimum assignment can be substantially higher than the experienced travel time in a user equilibrium assignment, and (2) the total travel time in user equilibrium can be substantially higher than the total travel time in system optimum. By seeking system optimal traffic flows subject to user constraints, a compromise assignment can be obtained that balances system and user objectives. To this aim, a linear model and an efficient heuristic algorithm are proposed in this paper. A computational study shows that the proposed model, along with the heuristic algorithm, is able to provide fair solutions with near-optimal total travel time within very short computational time.