Integral Reinforcement Reciprocal Zhang Neural Net for Time-Varying Linear Equations

Solving time-varying linear equations is a com-mon challenge encountered in engineering and science. The development of neural dynamic systems has led to the creation of various feedback neural nets that effectively solve continuous time-varying linear equations. In this context, reciprocal Zhang neural net (RZNN) was developed, which is focused by this work, providing an explicit inverse-free continuous model for solving time-varying linear equations. However, the continuous model often struggle with noise interference, e.g., Gaussian noise, during the solving process. This paper enhances RZNN model with an integral reinforcement term to improve robustness of the explicit continuous models with noisy conditions. Furthermore, we optimize the RZNN solving process, making it easier for implementation in circuit systems and numerical simulations.