Robust Pricing of Transportation Networks Under Uncertain Demand

Traditionally, tolls on transportation networks are determined on the basis of a single value of travel demand, deterministic elastic demand relationships, or informal scenario analysis. However, since the demand on the network cannot be forecast perfectly, pricing may prove to be suboptimal when the realized value of demand deviates significantly from the planned value. Therefore, there Is a need fora robust pricing scheme that accounts for demand uncertainty. Optimal pricing is examined through marginal costs in which origin-destination travel demand is a random variable to understand better the direct impact and sensitivity of the uncertainty. Three methods are evaluated for determining robust prices: inflation or deflation of the planning demand, averaging tolls front various planning demands, and genetic algorithms. The performance of these three methods is evaluated by analyzing user equilibrium for various future travel demand scenarios. From the results of the analysis, a more robust pricing scheme that accounts for variations in demand is developed.