An Application of Neural Network Optimization Approach

An approach based on Augmented Lagrange programming neural network has been proposed for solving a multi-reservoir power system problem. This technique is based on the Lagrange multiplier theory and search for solutions satisfying the necessary conditions of optimality in the state space. The equilibrium point of the network corresponds to the Lagrange solution of the problem and is asymptotically stable. Here the main objective is to determine the optimal amounts of water to be released from each reservoir during each interval of the scheduling horizon so as to minimize the expected production cost. The method takes into account the water transportation delays between upstream and downstream reservoirs. An example consisting of a multi-chain cascade of four reservoir type hydro-plants and interconnection lines to neighboring systems through which energy may be exchanged is solved with proposed optimizer, and the results are compared with those obtained using conventional conjugate gradient method. Key WordsConstrained optimization, hydro-power generation, energy function, augmented lagrange programming neural network optimizer, asymptotic stability Nomenclature = ) ( k E C Production cost in period k