Zero-shot Generalization in Inventory Management: Train, then Estimate and Decide

TL;DR

This paper introduces a unifying framework and a generalizable deep reinforcement learning policy for inventory management that performs well on unseen problem instances with unknown parameters, using a three-phase Train, Estimate, and Decide approach.

Contribution

It proposes the TED framework and GC-LSN policy for zero-shot generalization in inventory management, enabling effective decision-making under parameter uncertainty without retraining.

Findings

GC-LSN outperforms traditional policies when parameters are known.

GC-LSN combined with Kaplan-Meier estimator shows superior empirical results under uncertainty.

The framework effectively handles diverse inventory challenges with unknown demand and lead times.

Abstract

Deploying deep reinforcement learning (DRL) in real-world inventory management presents challenges, including dynamic environments and uncertain problem parameters, e.g. demand and lead time distributions. These challenges highlight a research gap, suggesting a need for a unifying framework to model and solve sequential decision-making under parameter uncertainty. We address this by exploring an underexplored area of DRL for inventory management: training generally capable agents (GCAs) under zero-shot generalization (ZSG). Here, GCAs are advanced DRL policies designed to handle a broad range of sampled problem instances with diverse inventory challenges. ZSG refers to the ability to successfully apply learned policies to unseen instances with unknown parameters without retraining. We propose a unifying Super-Markov Decision Process formulation and the Train, then Estimate and Decide (TED) framework to train and deploy a GCA tailored to inventory management applications. The TED framework consists of three phases: training a GCA on varied problem instances, continuously estimating problem parameters during deployment, and making decisions based on these estimates. Applied to periodic review inventory problems with lost sales, cyclic demand patterns, and stochastic lead times, our trained agent, the Generally Capable Lost Sales Network (GC-LSN) consistently outperforms well-known traditional policies when problem parameters are known. Moreover, under conditions where demand and/or lead time distributions are initially unknown and must be estimated, we benchmark against online learning methods that provide worst-case performance guarantees. Our GC-LSN policy, paired with the Kaplan-Meier estimator, is demonstrated to complement these methods by providing superior empirical performance.