Robust neural dynamics with adaptive coefficient applied to solve the dynamic matrix square root

Abstract Zeroing neural networks (ZNN) have shown their state-of-the-art performance on dynamic problems. However, ZNNs are vulnerable to perturbations, which causes reliability concerns in these models owing the potentially severe consequences. Although it has been reported that some possess enhanced robustness but cost worse convergence speed. In order address problems, a robust with an adaptive coefficient (RNDAC) model is proposed, aided by novel activation function and evolution formula boost speed preserve accuracy. validate analyze of RNDAC model, applied solve matrix square root (DMSR) problem. Related experiment results show reliably solves DMSR question perturbed various noises. Using we able reduce residual error from 10 $$^1$$ 1 $$^{-4}$$ - 4 noise reached satisfying competitive speed, converges within 3 s.