Network-optimized road pricing: Part I: A parable and a model

Part I of a two-part series, this paper recites a parable and formulates a stochastic optimization model that determines optimal link tolls on a road network whose users' value of time is a random variable. The parable, introducing the problem, demonstrates the importance of the variability of the value of time. The model, cast as a variational inequality, becomes a specialized form of a bicriterion user-equilibrium traffic assignment. Its solution is a set of efficient tolls for all links in the network. These tolls induce an equilibrium traffic flow that is at once system-optimal and user-optimal-for all trips, regardless of their value of time. Part II develops a solution algorithm, gives examples, and provides performance statistics.