Simple and Effective Re-Optimization for Revenue Management

What is it about?

Today's business problems are large-scale and complex. Solving the problem to optimality is often considered impossible, and many simply re-optimize a simple model periodically (e.g., airlines would re-optimize their seat-booking system several times in a month and some hotels would re-optimize their room-booking allocation at least once a week). While this approach is very practical, its performance relative to the optimal solution is not well-understood (e.g., how far are we from optimal?). This paper addresses this issue and shows that, in some settings, the performance of re-optimizations is indeed very close to optimal.

Featured Image

Why is it important?

This paper shows that sometimes a very simple solution to a very complicated problem exists. This has an important implication that, perhaps, many of the simple heuristics implemented in practice for tackling large-scale and complicated problems are not so naive after all!