Improved Zeroing Neural Network for Solving Time-Variant Linear Equation with Harmonic Noise

Time-variant linear equation (TVLE) is commonly involved in the science and engineering. Zeroing neural network (ZNN) and its variants have been developed and exhibited remarkable performance (e.g., noise rejection and finite-time convergence) on solving the TVLE. In this paper, we delve into a more in-depth investigation by enhancing the ZNN model for solving TVLE in the presence of harmonic noise. Such a model enables calculation error tend to zero, thereby facilitating the identification of the exact TVLE solution. To verify the supremacy of this enhanced ZNN model, we conduct a rigorous evaluation, assessing its properties such as convergence and robustness through various illustrative examples.