A pareto archive evolutionary strategy based radial basis function neural network training algorithm for failure rate prediction

This paper outlines a radial basis function neural network approach to predict the failures in overhead distribution lines of power delivery systems. The RBF networks are trained using historical data. The network sizes and errors are simultaneously minimized using the Pareto Archive Evolutionary Strategy algorithm. Mutation of the network is carried out by invoking an orthogonal least square procedure. The performance of the proposed method was compared to a fuzzy inference approach and with multilayered perceptrons. The results suggest that this approach outperforms the other techniques for the prediction of failure rates. Categories and Subject Descriptors I.2.6 [Artificial Intelligence]: Learning – connectionism and neural nets, parameter learning. General Terms: Algorithms

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: Grant Cochenour, Jerad Simon, Sanjoy Das, Anil Pahwa, Surasish Nag
• Keywords: Fuzzy Inference Approach, GECCO 2005, Neural Network Approach, Overhead Distribution Lines