Revenue management in the car rental industry: A stochastic programming approach

A stochastic programming approach for a network revenue management problem with flexible capacities is presented. The study focuses on a car rental network, which has the special property that the fleet distribution between rental stations in the network can easily be adjusted at determined costs. Our method simultaneously optimizes the fleet distribution on the network and the capacity controls at station level. A two-stage stochastic program is considered, where the demand uncertainty in the second stage is approximated by a finite number of scenarios. The performance of the proposed stochastic method is tested by simulation on a small car rental network and compared with the results of a deterministic program.