Application of artificial neural network to economic load dispatch

This paper provides a maiden attempt to explore the feasibility of using a multilayer feedforward neural network using the backpropagation and resilient propagation algorithms to solve the economic load dispatch problem for IEEE 14 and IEEE 30 bus systems. The effect of several important parameters such as the number of hidden units, type of activation function, normalisation, the learning rate parameter, the optimum number of training patterns, etc., in training the neural network has been clearly brought out