Hopfield neural network approach for multi-reservoir network control

In this paper a novel approach based on linear programming neural network (LPNN) has been proposed for controlling the interconnected multi-reservoir systems. The objective here is to maximize the overall energy generation benefit subject to a number of hydro system constraints such as bounds on reservoir releases and storages. The method is based on the solution of the ordinary differential equations obtained from the energy function of the control problem. The developed method has been implemented on the interconnected four reservoir standard system. The devised technique has shown the potential to achieve true optimal results as compared with other popular techniques.