|
Abstract
|
Hopfield neural network (HNN) is considered as an artificial model derived from the brain structures and it is an important model
that admits an adequate performance in neurocomputing. In this article, we solve a dynamical model of 3D HNNs via
Atangana–Baleanu (AB) fractional derivatives. To find the numerical solution of the considered dynamical model, the well-known
Predictor-Corrector (PC) method is used. A number of cases are taken by using two different sets of values of the activation
gradient of the neurons as well as six different initial conditions. )e given results have been perfectly established using the
different fractional-order values on the given derivative operator. )e objective of this research is to investigate the dynamics of
the proposed HNN model at various values of fractional orders. Nonlocal characteristic of the AB derivative contains the memory
in the system which is the main motivation behind the proposal of this research.
|