“Dynamic Pricing and Inventory Control with Nonparametric Demand Learning” by Prof. Xiuli CHAO
Professor Xiuli CHAO
Department of Industrial and Operations Engineering
College of Engineering
The University of Michigan
We consider a retailer selling a single nonperishable product over a finite horizon. Demand is stochastic and price-dependent. At the beginning of each period, the firm determines its selling price and inventory replenishment decisions, but it knows neither the dependency of demand on selling price nor the distribution of demand uncertainty, hence it has to make pricing and ordering decisions only based on historical demand data. We propose a nonparametric data-driven policy that learns the demand-price relationship and the random error distribution on the fly. The policy integrates the phases of exploration and exploitation and converges to the true optimal solution. Besides convergence of optimal policies, we also establish the convergence rate of the regret, defined as the profit loss compared with that of the optimal solution when the firm had known the random demand information. This is joint work with Beryl Chen and Hyun-Soo Ahn.