Smart Grid Builder: AI agents for clean energy

Описание: Smart-Grid is a multi-agent reinforcement learning system for optimised power dispatch in the electric grid.The simulator treats the components of an electrical grid (producers, consumers, storage/batteries and grid operators) as reinforcement agents and uses curriculum learning to learn the optimum approach to integrate renewable energy resources like solar/wind power with the grid. The simulations can be used to study good and bad decision making in such systems. Ex:
Understand how a battery should optimally perform when the supply is overwhelmed by demand.

Github: https://github.com/haran2001/open-sma...

Blog:   / smart-grid-multi-agent-reinforcement-learn...