Large-scale economic dispatch using an improved Hopfield neural network

The application of Hopfield neural networks to the solution of large-scale economic-dispatch (ED) problems is proposed. A special methodology for improving the performance of Hopfield neural networks for solving combinatorial-optimisation problems has already been published. That method is quicker and more accurate. Accordingly, Gee's Hopfield neural networks (GHN) have been used instead of standard Hopfield neural networks. Gee's method has been modified to solve ED. Constraints are handled using a combination of Gee's model and Abe's model. An efficient simulation algorithm is discussed for solving the economic-dipatch problem. The proposed method (PHN) has achieved efficient and accurate solutions for power systems with 3, 20, 40, 80 , 120, 160 and 240 units. The PHN results are very close to those obtained using classical numerical methods.