Garrett J. van Ryzin

Simulation-Based Optimization of Virtual Nesting Controls for Network Revenue Management

Virtual nesting is a popular control strategy for network revenue management. Bertsimas and de Boer recently proposed a simulation-based optimization method for computing the protection levels in a virtual nesting control scheme. In this talk, we analyze a variation of the Bertsimas-de Boer method that is computationally more efficient and has provable convergence guarantees. The approach is based on a continuous model of capacity and demand that allows for simple and exact subgradient computations. In several test problems on realistic networks, the method produces significant performance improvements relative to the protection levels produced by heuristic virtual nesting schemes.
[This is joint work with Gustavo Vulcano, New York University.]