English

Markdown Pricing Under an Unknown Parametric Demand Model

Machine Learning 2023-12-27 v1 Computer Science and Game Theory

Abstract

Consider a single-product revenue-maximization problem where the seller monotonically decreases the price in nn rounds with an unknown demand model coming from a given family. Without monotonicity, the minimax regret is O~(n2/3)\tilde O(n^{2/3}) for the Lipschitz demand family and O~(n1/2)\tilde O(n^{1/2}) for a general class of parametric demand models. With monotonicity, the minimax regret is O~(n3/4)\tilde O(n^{3/4}) if the revenue function is Lipschitz and unimodal. However, the minimax regret for parametric families remained open. In this work, we provide a complete settlement for this fundamental problem. We introduce the crossing number to measure the complexity of a family of demand functions. In particular, the family of degree-kk polynomials has a crossing number kk. Based on conservatism under uncertainty, we present (i) a policy with an optimal Θ(log2n)\Theta(\log^2 n) regret for families with crossing number k=0k=0, and (ii) another policy with an optimal Θ~(nk/(k+1))\tilde \Theta(n^{k/(k+1)}) regret when k1k\ge 1. These bounds are asymptotically higher than the O~(logn)\tilde O(\log n) and Θ~(n)\tilde \Theta(\sqrt n) minimax regret for the same families without the monotonicity constraint.

Keywords

Cite

@article{arxiv.2312.15286,
  title  = {Markdown Pricing Under an Unknown Parametric Demand Model},
  author = {Su Jia and Andrew Li and R. Ravi},
  journal= {arXiv preprint arXiv:2312.15286},
  year   = {2023}
}
R2 v1 2026-06-28T14:00:45.646Z