Predefined‐Time Convergent Zeroing Neural Network for Time‐Varying Quadratic Program in the Presence of Noise

ABSTRACT Quadratic programming (QP) is fundamental across science and engineering, and numerous algorithms have been developed to compute QP solutions. To advance this line of research, this paper proposes a new zeroing neural network (ZNN) for time‐varying QP (TVQP) subject to noise. The proposed ZNN employs a tailored nonlinear activation and incorporates integral augmentation, achieving predefined‐time convergence and strong noise tolerance. Rigorous theoretical analysis, together with simulation studies, corroborates the effectiveness and robustness of the proposed model.