Application of fuzzy dynamic programming and neural network in generation scheduling

This paper introduces a unit commitment method based on artificial neural network (ANN) and fuzzy dynamic programming (FDP). Comparison among two ANN methods and FDP method is discussed. The experimental results indicate that the proposed ANN algorithm can significantly reduce the execution time in unit commitment based on binary codes, and the application of Gray code can reduce the dimensions of neural network and its training time consequently. The fuzzy approach to unit commitment achieves a reasonable operation cost and optimum state for constrained power systems.