Reinforcement Learning with Probabilistic Boolean Network Models of Smart Grid Devices

The area of smart power grids needs to constantly improve its efficiency and resilience, provide high quality electrical in a resilient grid, while managing faults avoiding failures. Achieving this requires component reliability, adequate maintenance, studied failure occurrence. Correct system operation involves those activities novel methodologies detect, classify, isolate failures model simulate processes with predictive algorithms analytics (using data analysis asset condition plan perform activities). In paper, we showcase the application complex-adaptive, self-organizing modeling method, Probabilistic Boolean Networks (PBNs), as way towards understanding dynamics grid devices, characterize their behavior. This work demonstrates that PBNs are equivalent standard Reinforcement Learning Cycle, which agent/model has an interaction environment receives feedback from it form reward signal. Different structures were created preferred information can be used guide PBN avoid fault conditions