Simulation-Based Benchmarking of Reinforcement Learning Agents for Personalized Retail Promotions

TL;DR

This paper introduces a simulation framework for benchmarking reinforcement learning agents in retail coupon targeting, demonstrating that advanced RL methods outperform static policies in sparse reward environments.

Contribution

It presents a comprehensive simulation platform for retail AI benchmarking and evaluates RL agents trained on offline data for personalized promotions.

Findings

Contextual bandit and deep RL outperform static policies.

Less over-fitting in RL methods improves performance.

Simulation framework accelerates retail AI development.

Abstract

The development of open benchmarking platforms could greatly accelerate the adoption of AI agents in retail. This paper presents comprehensive simulations of customer shopping behaviors for the purpose of benchmarking reinforcement learning (RL) agents that optimize coupon targeting. The difficulty of this learning problem is largely driven by the sparsity of customer purchase events. We trained agents using offline batch data comprising summarized customer purchase histories to help mitigate this effect. Our experiments revealed that contextual bandit and deep RL methods that are less prone to over-fitting the sparse reward distributions significantly outperform static policies. This study offers a practical framework for simulating AI agents that optimize the entire retail customer journey. It aims to inspire the further development of simulation tools for retail AI systems.