A B C
Z. Naturforsch. 69a, 129 – 134 (2014)
doi:10.5560/ZNA.2013-0082
Optimal Range of Parameters of Hopfield's Neural Network for Shortest Path Computation in Routing
Farah Sarwar1 and Shaukat Iqbal2
1 Department of Computer Science, COMSATS Institute of Information Technology, Sahiwal, Pakistan
2 Department of Computer Science, COMSATS Institute of Information Technology, Vehari, Pakistan
Received July 18, 2013 / revised November 4, 2013 / published online January 22, 2014
Reprint requests to: S. I.; E-mail: shaukat.iqbal.k@gmail.com
This presentation provides a mathematical survey of parameters used to adjust weights in Hopfield neural network (HNN). These parameter values are used to adjust the rate of convergence, optimality ratio, and precision of the network for a given problem. Optimal shortest routing path computation is discussed here as one of the optimization problems. A detailed analysis of one such energy function is carried out by varying the parameter values one by one in different ranges. A comprehensive conclusion is presented here about the proper criterion for selecting the values of these effectual parameters.
Key words: Hopfield Neural Network; Optimal Parameters; Shortest Path Computation; Routing.
Full-text PDF