Contextual Dynamic Pricing: Algorithms, Optimality, and Local Differential Privacy Constraints

2024-06-04

arXiv.org

AI Summary

This paper studies contextual dynamic pricing under local differential privacy (LDP) constraints, where the demand follows an unknown generalized linear model, aiming to minimize cumulative regret relative to a clairvoyant policy. Methodologically, it establishes the first theoretical connection between dynamic pricing and contextual multi-armed bandits; proposes the first online pricing framework supporting mixed privacy constraints; and integrates confidence-bound estimation, exploration–exploitation switching, and stochastic gradient updates. Theoretically, it proves that $sqrt{dT}$ is the minimax-optimal regret rate, derives a tight LDP-specific bound of $dsqrt{T}/varepsilon$, characterizes the privacy cost, and achieves Pareto-improvement in the privacy–utility trade-off. The algorithm matches the minimax lower bound, and empirical evaluation on real-world datasets demonstrates significant improvements over existing baselines.

Technology Category

• Machine Learning: Privacy
• Search and Optimization: Learning to Search
• Game Theory and Economic Paradigms: Mechanism Design

Application Category

• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Security and Privacy: Privacy-enhancing technologies

Abstract

We study contextual dynamic pricing problems where a firm sells products to $T$ sequentially-arriving consumers, behaving according to an unknown demand model. The firm aims to minimize its regret over a clairvoyant that knows the model in advance. The demand follows a generalized linear model (GLM), allowing for stochastic feature vectors in $mathbb R^d$ encoding product and consumer information. We first show the optimal regret is of order $sqrt{dT}$, up to logarithmic factors, improving existing upper bounds by a $sqrt{d}$ factor. This optimal rate is materialized by two algorithms: a confidence bound-type algorithm and an explore-then-commit (ETC) algorithm. A key insight is an intrinsic connection between dynamic pricing and contextual multi-armed bandit problems with many arms with a careful discretization. We further study contextual dynamic pricing under local differential privacy (LDP) constraints. We propose a stochastic gradient descent-based ETC algorithm achieving regret upper bounds of order $dsqrt{T}/epsilon$, up to logarithmic factors, where $epsilon>0$ is the privacy parameter. The upper bounds with and without LDP constraints are matched by newly constructed minimax lower bounds, characterizing costs of privacy. Moreover, we extend our study to dynamic pricing under mixed privacy constraints, improving the privacy-utility tradeoff by leveraging public data. This is the first time such setting is studied in the dynamic pricing literature and our theoretical results seamlessly bridge dynamic pricing with and without LDP. Extensive numerical experiments and real data applications are conducted to illustrate the efficiency and practical value of our algorithms.

Problem

• Optimizing dynamic pricing under unknown demand model
• Achieving minimal regret with GLM demand
• Balancing privacy constraints in pricing algorithms

Innovation

• Generalized linear model for dynamic pricing
• Confidence bound and ETC algorithms
• Stochastic gradient descent under LDP