Newsvendor Model with Deep Reinforcement Learning

TL;DR

This paper introduces a deep reinforcement learning approach using a Twin-Delayed Deep Deterministic Policy Gradient agent to solve the Newsvendor problem with varying demand distributions, achieving optimal policies aligned with analytical solutions.

Contribution

It presents a novel RL-based method for the Newsvendor model that adapts to changing demand distributions across different days, demonstrating learned optimal behavior.

Findings

RL agent learns optimal policies consistent with analytical solutions

Agent identifies and adapts to different demand distributions

Method handles complex, realistic demand scenarios

Abstract

I present a deep reinforcement learning (RL) solution to the mathematical problem known as the Newsvendor model, which seeks to optimize profit given a probabilistic demand distribution. To reflect a more realistic and complex situation, the demand distribution can change for different days of the week, thus changing the optimum behavior. I used a Twin-Delayed Deep Deterministic Policy Gradient agent (written as completely original code) with both an actor and critic network to solve this problem. The agent was able to learn optimal behavior consistent with the analytical solution of the problem, and could identify separate probability distributions for different days of the week and behave accordingly.