Fusion of soft computing and hard computing techniques for power energy system

A power energy system normally consists of a number of power plants distributed over a wide area, connected by an electric power grid. Optimal operation of such a system involves the short-term load forecasting problem, the unit commitment problem and the start-up scheduling problem. All these problems are highly nonlinear and the optimization problems contain a number of local optima, making them difficult to solve by the conventional "hard computing" techniques alone. Thus, an approach based on a fusion of different types of techniques is required. This paper presents an overview of a fusion of soft computing and hard computing techniques, which the authors have developed so far aiming at achieving an optimal load dispatch and automatic power plant start-up control system.