Home Energy Management with V2X Capability using Reinforcement Learning

Abstract

The increased demand for Smart Home control technologies and the rapid growth of AI-based approaches provide an opportunity to develop systems that significantly reduce homeowners’ electricity costs and decrease the inconvenience of power outages. Reinforcement Learning is an AI tool for training systems to perform complex tasks. The paper proposes an RL-based Home Energy Management System to optimally manage a user’s electricity cost while maximizing user comfort and convenience. The system can control the smart home in the presence of the uncertainty and variability of Solar power generation and a varying electricity demands of a homeowner.