Online Revenue Maximization with Unknown Concave Utilities

TL;DR

This paper investigates online revenue maximization with unknown consumer distributions and concave utilities, proposing algorithms with sub-linear regret bounds under certain smoothness conditions.

Contribution

It introduces a framework for online pricing with unknown, non-linear, concave utilities, and provides two algorithms with provable regret guarantees.

Findings

Algorithms achieve regret bounds of O(T^{2/3} (log T)^{1/3}) and O(T^{1/2} (log T)^{1/2})

Imposing third-order smoothness enables concavity in the optimization problem

Addresses the challenge of unknown consumer utilities in online revenue maximization

Abstract

We study an online revenue maximization problem where the consumers arrive i.i.d from some unknown distribution and purchase a bundle of products from the sellers. The classical approach generally assumes complete knowledge of the consumer utility functions, while recent works have been devoted to unknown linear utility functions. This paper focuses on the online posted-price model with unknown consumer distribution and unknown consumer utilities, given they are concave. Hence, the two questions to ask are i) when is the seller's online maximization problem concave, and ii) how to find the optimal pricing strategy for non-linear utilities. We answer the first question by imposing a third-order smoothness condition on the utilities. The second question is addressed by two algorithms, which we prove to exhibit the sub-linear regrets of $O(T^{\frac{2}{3}} (\log T)^{\frac{1}{3}})$ and $O(T^{\frac{1}{2}} (\log T)^{\frac{1}{2}})$ respectively.