Reinforcement Learning approach for safe and fair coordination of distributed energy resources

This thesis offers a comprehensive and practical solution for achieving safe and fair coordination of DERs in distribution networks. Our proposed approaches combine model-free characteristics, fairness, decentralization, and safety within a unified framework that addresses the realities of low network visibility, limited communication, and privacy concerns. The findings provide insights for researchers, system operators, and policymakers seeking to facilitate equitable and resilient energy transitions in power systems.