Distributional constrained reinforcement learning for supply chain optimization
This work addresses the challenge of safe and predictable reinforcement learning for supply chain management, representing an incremental improvement over existing constrained policy optimization methods.
The paper tackles the problem of unreliable constraint satisfaction in reinforcement learning for supply chain optimization by introducing Distributional Constrained Policy Optimization (DCPO), which improves convergence rates and ensures reliable constraint satisfaction while reducing variance in returns.
This work studies reinforcement learning (RL) in the context of multi-period supply chains subject to constraints, e.g., on production and inventory. We introduce Distributional Constrained Policy Optimization (DCPO), a novel approach for reliable constraint satisfaction in RL. Our approach is based on Constrained Policy Optimization (CPO), which is subject to approximation errors that in practice lead it to converge to infeasible policies. We address this issue by incorporating aspects of distributional RL into DCPO. Specifically, we represent the return and cost value functions using neural networks that output discrete distributions, and we reshape costs based on the associated confidence. Using a supply chain case study, we show that DCPO improves the rate at which the RL policy converges and ensures reliable constraint satisfaction by the end of training. The proposed method also improves predictability, greatly reducing the variance of returns between runs, respectively; this result is significant in the context of policy gradient methods, which intrinsically introduce significant variance during training.