Thanos Avramidis (University of Montréal)

Forecast Effects on Revenue Management

Two leading approaches in the practice of network revenue management are deterministic demand modeling with linear programming optimization
versus stochastic demand modeling with nonlinear programming optimization. We compare the performance of these techniques on small network problems by a designed experiment where the factors of interest are forecast accuracy and forecast error distribution (both observable by the provider). We also study the effects of competition on the demand forecasts and the expected revenues via experiments with two competing providers, each practicing revenue management.