A parametric iteration method of stochastic dynamic programming for optimal dispatch of hydroelectric plants

In this paper the zero points of Chebyshev´s polynomials are used for approximating the benefit functions of the rest stage of the stochastic dynamic programming method. It has been proved that the method is an efficient way for overcoming too many combination states of the stochastic dynamic programming and improving the conventional discrete dynamic programming algorithm because of requiring less computer memory and CPU time. The method is used to solve the optimal dispatching scheme of two cascade hydroelectric stations with pluriennial regulating reservoirs, considering the inflows´ stochastic property, their mutual dependence in space and the reliability of the system operation. The results obtained are satisfactory