Perturbed Pricing

Abstract

We propose a simple randomized rule for the optimization of prices in revenue management with contextual information. It is known that the certainty equivalent pricing rule, albeit popular, is sub-optimal. We show that, by allowing a small amount of randomization around these certainty equivalent prices, the benefits of optimal pricing and low regret are achievable.