Application of a genetic algorithm in an artificial neural network to calculate the resonant frequency of a tunable single-shorting-post rectangular-patch antenna

✍ Scribed by Shyam S. Pattnaik; Bonomali Khuntia; Dhruba C. Panda; Dipak K. Neog; S. Devi; Malay Dutta

• Publisher: John Wiley and Sons
• Year: 2004
• Tongue: English
• Weight: 82 KB
• Volume: 15
• Category: Article
• ISSN: 1096-4290
• DOI: 10.1002/mmce.20060

✦ Synopsis

In this article, an efficient application of a genetic algorithm (GA) in an artificial neural network (ANN) to calculate the resonant frequency of a coaxially-fed tunable rectangular microstrip-patch antenna is presented. For a normal feed-forward back-propagation algorithm, with a compromise between time and accuracy, it is difficult to train the network to achieve an acceptable error tolerance. The selection of suitable parameters of ANNs in a feed-forward network leads to a high number of man-hours necessary to train a network efficiently. However, in the present method, the GA is used to reduce the man-hours while training a neural network using the feed forward-back-propagation algorithm. It is seen that the training time has also been reduced to a great extent while giving high accuracy. The results are in very good agreement with the experimental results.