Robust stability analysis of cohen-grossberg neural networks with time delays

In the scope of this thesis, the global asymptotic robust stability properties of Cohen-Grossberg neural networks with multiple time delays were analyzed. In this neural network model, it is assumed that neuron activations belong to the class of non-decreasing slope-bounded nonlinear functions. In addition, it is considered that the network parameters in the system have bounded upper norms. Under these assumptions, firstly, by using the homeomorphism theorem, new sufficient conditions that ensure the existence and uniqueness of the equilibrium point are obtained independently of the time delay and only depending on the network parameters. Secondly, by using Lyapunov stability theorems, it has been shown that the conditions that the obtained new sufficient conditions also realize the global asymptotic robust stability of the equilibrium point of the system. Finally, some comparisons were made using numerical examples in order to show the superiority of these conditions over the results obtained under similar assumptions in the literature.