Stability analysis of quaternion-valued neural networks with both discrete and distributed delays

摘要

The existence, uniqueness and stability of the equilibrium of quaternion-valued neural networks (QVNNs) with both discrete and distributed delays are investigated in this paper. The considered model is managed as a single entirety without decomposition. Based on homeomorphic mapping theorem and linear matrix inequality, several sufficient criteria are derived to ascertain the aforementioned QVNNs to be globally asymptotically stable and exponentially stable. Moreover, provided criteria can be verified by the linear matrix inequality (LMI) toolbox in MATLAB. Finally, one simulation example is demonstrated to verify the effectiveness of obtained results.