Hopfield networks with jump Markov parameters
Adrian‐Mihail Stoica, I. Yaesh · 2005
Hopfield networks are symmetric recurrent neural networks which exhibit motions in the state space which converge to minima of energy. Hopfield networks can be used to solve practical complex problems such as implement associative memory, linear programming solvers and optimal guidance problems. In such practical problems, the Hopfield network, may be subject to disturbance signals which can be modelled as finite energy signals. In this paper, we adopt the Lur'e-Postnikov systems approach to analyze Hopfield networks and suggest a training algorithm leading to minimum L2 gain from the disturbance signals to the error of the network with respect and to its equilibrium points. The suggested algorithm is applied to a numerical example from the field of magnetic heading determination.