Grida: introducing a self-learning artificial intelligence for autonomous network planning

Abstract

This paper presents a self-learning artificial intelligence for grid planning. The self-learning means that no expert input data is required, which is a large advantage in electricity distribution planning with little expert data for optimal grids. The performance of the self-learning agent is comparable to existing grid design when compared to real-life grids of a festival of 7.000 people and a city center of 26.000 people.