2024 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm) – Workshop on Safe Reinforcement Learning for Smart Grid Control and Operations
CALL FOR PAPERS
Power systems are complex cyber-physical systems requiring sophisticated control and decision-making capabilities for reliable, efficient, and secure operations. Ensuring safety is critical, given the potential for widespread disruptions if operational constraints are violated in this critical infrastructure. Reinforcement learning (RL) presents a promising data-driven approach to developing intelligent control policies, but deploying RL necessitates new methods that can respect safety constraints through safe exploration, constrained optimization, and stability guarantees.
Smart grids are rapidly evolving with increasing renewable energy, distributed resources, and data-driven automation. While RL enables more intelligent and adaptive grid control, the domain's strict operational constraints and safety requirements necessitate tailored, safe RL techniques. This emerging interdisciplinary area combining RL, control theory, and power systems is gaining traction but lacks comprehensive dedicated treatments.
To bring together the diverse community of Safe RL and smart grid researchers together to steer discussions on challenges and perspectives of using RL for smart grid applications, we propose the Safe RL Workshop for Smart Grid Control and Operations. The sessions will focus on the progress and challenges in safe RL, emphasizing its application to the complex cyber-physical systems of power grids. Discussions will delve into topics such as safety constraint satisfaction during exploration, optimization under constraints, stability guarantees, and the operational challenges of deploying RL in smart grids where reliability and security are extremely important. By concentrating on these areas, the workshop aims to highlight the unique requirements and solutions for safe RL deployment in power systems, moving towards the goal of achieving efficient, secure, and resilient grid operations.
TOPICS OF INTEREST
With the notion of safety in mind, we encourage submissions on topics including but not limited to:
- Safe exploration in electrical power systems
- Definitions and metrics for safety in smart grid control
- Safe RL in Renewable Energy Integration and Demand Response
- Integrating safety constraints into reinforcement learning policies for power systems
- Ensuring operational safety in non-stationary smart grid conditions
- Off-policy and offline RL for safety in smart grid operations
- Risk assessment and robust decision-making frameworks for grid stability and security
- Case studies on the application of safe RL in smart grid operations and control
- Ethical considerations and social impact of implementing RL in power systems
We invite submissions of abstracts up to 4 pages, excluding references that discuss early-stage or mature work related to the topics above. Accepted abstracts will be presented as spotlight talks within thematic sessions at the workshop.
IMPORTANT DATES
- Paper submission deadline (regular and short papers): June 25, 2024
- Review results announced: July 10, 2024
- Submission of camera-ready papers for pre-proceedings: August 9, 2024
WORKSHOP CHAIRS
- Ming Jin, Bradley Department of Electrical and Computer Engineering at Virginia Tech, Email: jinming@vt.edu
- Javad Lavaei, Department of Industrial Engineering and Operations Research at UC Berkeley, Email: lavaei@berkeley.edu
- Ali Mehrizi-Sani, Virginia Tech, Email: mehrizi@vt.edu
- Shangding Gu, UC Berkeley, Email: shangding.gu@tum.de
