Application of a composite differential evolution algorithm in optimal neural network design for propagation path-loss prediction in mobile communication systems

Sotirios P. Sotiroudis, Sotirios K. Goudos, Konstantinos A. Gotsis, Katherine Siakavara, John N. Sahalos

Research output: Contribution to journalArticle

32 Citations (Scopus)

Abstract

In this letter, we present an alternative procedure for the prediction of propagation path loss in urban environments, which is based on artificial neural networks (ANNs). The correct selection of a neural network size can increase its response speed and therefore increase the overall system performance. We apply a recently proposed Differential Evolution (DE) algorithm, namely the Composite DE (CoDE) in order to design an optimal ANN for path-loss propagation prediction. CoDE uses three different trial-vector generation strategies with three preset control parameter settings. We compare CoDE with other popular DE strategies. We present two different ANN design cases with two and three hidden layers, respectively. The general performance of both the ANNs shows their effectiveness to yield results with satisfactory accuracy in short time. The received results are compared to the respective ones yielded by the ray-tracing model and exhibit satisfactory accuracy.

Original languageEnglish
Article number6476630
Pages (from-to)364-367
Number of pages4
JournalIEEE Antennas and Wireless Propagation Letters
Volume12
DOIs
Publication statusPublished - 2013

Keywords

  • Differential evolution (DE)
  • evolutionary algorithms (EAs)
  • mobile communications
  • neural network
  • optimization methods
  • propagation path loss
  • self-adaptive differential evolution

Fingerprint Dive into the research topics of 'Application of a composite differential evolution algorithm in optimal neural network design for propagation path-loss prediction in mobile communication systems'. Together they form a unique fingerprint.

  • Cite this