Dual-Agent Deep Reinforcement Learning for Dynamic Pricing and Replenishment

TL;DR

This paper introduces a dual-agent deep reinforcement learning framework for dynamic pricing and inventory replenishment, effectively handling decision frequency discrepancies and demand complexities to optimize profits.

Contribution

It presents a novel dual-agent DRL algorithm with a two-timescale scheme for joint pricing and replenishment under complex demand models.

Findings

The concavity of the profit function is established.

The integrated machine learning demand model improves accuracy.

Numerical results show the method's effectiveness in various scenarios.

Abstract

We study the dynamic pricing and replenishment problems under inconsistent decision frequencies. Different from the traditional demand assumption, the discreteness of demand and the parameter within the Poisson distribution as a function of price introduce complexity into analyzing the problem property. We demonstrate the concavity of the single-period profit function with respect to product price and inventory within their respective domains. The demand model is enhanced by integrating a decision tree-based machine learning approach, trained on comprehensive market data. Employing a two-timescale stochastic approximation scheme, we address the discrepancies in decision frequencies between pricing and replenishment, ensuring convergence to local optimum. We further refine our methodology by incorporating deep reinforcement learning (DRL) techniques and propose a fast-slow dual-agent DRL algorithm. In this approach, two agents handle pricing and inventory and are updated on different scales. Numerical results from both single and multiple products scenarios validate the effectiveness of our methods.