Competitive Robust Dynamic Pricing in Continuous Time with Fixed Inventories

TL;DR

This paper extends robust dynamic pricing models to continuous time with fixed inventories, using a variational inequality approach and fixed-point algorithms to ensure superior worst-case performance in uncertain demand environments.

Contribution

It develops a continuous-time robust pricing framework with a generalized Nash equilibrium and variational inequality formulation, incorporating explicit time lags and decision rules.

Findings

Robust pricing outperforms nominal pricing in worst-case scenarios.

Existence of a generalized robust Nash equilibrium under regularity conditions.

Efficient solution via fixed-point algorithms for the variational inequality.

Abstract

The problem of robust dynamic pricing of an abstract commodity, whose inventory is specified at an initial time but never subsequently replenished, originally studied by Perakis and Sood (2006) in discrete time, is considered from the perspective of continuous time. We use a multiplicative demand function to model the uncertain demand, and develop a robust counterpart to replace the uncertain demand constraint. The sellers' robust best response problem yields a generalized Nash equilibrium problem, which can be formulated as an equivalent, continuous-time quasi-variational inequality. We demonstrate that, for appropriate regularity conditions, a generalized robust Nash equilibrium exists. We show that the quasi-variational inequality may be replaced by an equivalent variational inequality, and use a fixed-point algorithm to solve the variational inequality. We also demonstrate how explicit time lags associated with price updating in real-world decision environments, as well as specific pricing decision rules, may be introduced to create a dual time scale formulation and the associated solutions computed. We illustrate, via numerical examples, how robust pricing based on our DPFI formulation offers generally superior and never inferior worst case performance compared to nominal pricing.