OPTIMIZATION AND LEARNING OF LOAD RESTORATION STRATEGIES

This paper describes art application of optimization and machine learn ing to load restoration in a generation-transmission system. An optimi zation procedure, combining a genetic algorithm and a power system dyn amic simulator, generates the appropriate sequence of operations for e ach state of the power system. A machine learning technique (induction of decision trees) is applied to extract decision criteria that will guide the load restoration after a generalized black-out. The paper al so presents the results of applying these techniques to a power system of realistic size. (C) 1997 Elsevier Science Ltd.