Reinforcement Learning for the Unit Commitment Problem

TL;DR

This paper applies reinforcement learning to the day-ahead unit commitment problem, formulating it as an MDP and developing algorithms that significantly reduce operational costs and computation time compared to traditional methods.

Contribution

The paper introduces three reinforcement learning algorithms tailored for the unit commitment problem and demonstrates their effectiveness over existing approaches.

Findings

27% reduction in operation costs

Significantly faster computation time (2.5 minutes vs. 2.5 hours)

Outperforms simulated annealing-based methods

Abstract

In this work we solve the day-ahead unit commitment (UC) problem, by formulating it as a Markov decision process (MDP) and finding a low-cost policy for generation scheduling. We present two reinforcement learning algorithms, and devise a third one. We compare our results to previous work that uses simulated annealing (SA), and show a 27% improvement in operation costs, with running time of 2.5 minutes (compared to 2.5 hours of existing state-of-the-art).