Technical Note—Fairness-Aware Online Price Discrimination with Nonparametric Demand Models.

The article focuses on integrating fairness constraints into dynamic online price discrimination using nonparametric demand models. It proposes a fairness-aware pricing framework that enforces relative price fairness, limiting price disparities between customer groups based on their unconstrained optimal prices, controlled by a parameter λ. The authors establish a novel regret lower bound of order Ω(T^4/5) for any fairness-aware pricing policy, reflecting the intrinsic difficulty of balancing revenue optimization and fairness under nonparametric demand uncertainty. They also develop an explore-and-exploit algorithm achieving a matching upper bound on regret, demonstrating the fundamental trade-offs introduced by relative fairness constraints compared to traditional dynamic pricing. The work advances theoretical understanding of ethical pricing in data-driven markets and offers methodological tools potentially applicable to other constrained learning problems.