A genetic based Service Restoration Algorithm for real-time operated Medium Voltage Distribution Networks using High Performance Computing

This paper presents a new Service Restoration Algorithm for solving the restoration problem of a Medium Voltage Network after a fault. The algorithm is oriented to solve real complex problems dealing with up to thousands of switches. The resulting huge combinatorial problem is solved combining a classical genetic based approach with a probabilistic driven mutation, a smart selection of switches to try, and some heuristic processes to solve constrains when needed. Moreover, performance requirements for real-time operations or a future more sophisticated mutilobjective function are achieved by a High Performance Computing design.