Journal of Beijing University of Posts and Telecommunications

  • EI核心期刊

JOURNAL OF BEIJING UNIVERSITY OF POSTS AND TELECOM ›› 2014, Vol. 37 ›› Issue (1): 90-93.doi: 10.13190/j.jbupt.2014.01.020

• Reports • Previous Articles     Next Articles

A Neural Network Based Method to Assess Electric Power Communication Network Risk

QI Feng1, LI Qi1, HAN Qian1, DU Yi2, QIU Xue-song1   

  1. 1. State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. China Electronic Systems Engineering Corporation, Beijing 100840, China
  • Received:2013-04-07 Online:2014-02-28 Published:2014-01-07

Abstract:

An improved back propagation neural network algorithm based on dichotomy was proposed for assessment of electric power communication network risk. The dichotomy was used to adjust the learning rate in the training process. It helps to change the learning rate automatically to the direction of optimization. Simulation shows that the improved algorithm's performance is optimized, such as convergence rate、error accuracy and training time.

Key words: back propagation neural network, learning rate, dichotomy, electric power communication network, assessment of risk

CLC Number: