Using Meta Reinforcement Learning to Bridge the Gap between Simulation and Experiment in Energy Demand Response

TL;DR

This paper introduces a meta-reinforcement learning approach to improve energy demand response experiments by efficiently leveraging simulated data, reducing the need for costly real-world training data.

Contribution

It applies meta-learning to bridge the simulation-reality gap in energy demand response, enhancing sample efficiency in real-world experiments.

Findings

Meta-learning improves learning efficiency in demand response tasks.

Simulated tasks effectively warm start real experiments.

Higher complexity tasks lead to better learning outcomes.

Abstract

Our team is proposing to run a full-scale energy demand response experiment in an office building. Although this is an exciting endeavor which will provide value to the community, collecting training data for the reinforcement learning agent is costly and will be limited. In this work, we apply a meta-learning architecture to warm start the experiment with simulated tasks, to increase sample efficiency. We present results that demonstrate a similar a step up in complexity still corresponds with better learning.