Adaptive Zeroing-Type Neural Dynamics for Solving Quadratic Minimization and Applied to Target Tracking

In conclusion, for general QM problems, the traditional approach solves them by employing some numerical or iterative algorithms [4]. A message-passing scheme for solving QM problems is presented in [5] by Ruozzi and Tatikonda. In addition, Zhang et al. present a QM-based dual-arm cyclic-motion-generation manipulator control scheme and analyze its properties from the perspective of cybernetics [3]. It is worth noting that although considerable research has been devoted to solving conventional QM problems, studies aimed explicitly at the time-varying quadratic minimization (TVQM) problem are insufficient. Traditional solutions have serious lag errors when facing large-scale time-varying issues, resulting in inadequate solution accuracy and even the collapse of the solution system [6]. To break through the dilemma that traditional algorithms cannot effectively deal with time-varying problems, Zhang et al. design the original zeroing neural network (OZNN) model [7]. The OZNN model employs the derivative information of the time-varying problem to predict its evolution direction and continuously adjust the solution strategy of the solution system through a named evolution function [8].