Title: Reinforcement Learning for Electricity Network Operation

Abstract: This paper presents the background material required for the Learning to Run Power Networks Challenge. The challenge is focused on using Reinforcement Learning to train an agent to manage the real-time operations of a power grid, balancing power flows and making interventions to maintain stability. We present an introduction to power systems targeted at the machine learning community and an introduction to reinforcement learning targeted at the power systems community. This is to enable and encourage broader participation in the challenge and collaboration between these two communities.