Noise-suppressing zeroing neural network for online solving time-varying nonlinear optimization problem: a control-based approach

Time-varying nonlinear optimization problems with different noises often arise in the fields of scientific and engineering research. Noises are unavoidable in the practical workspace, but the most existing models for time-varying nonlinear optimization problems carry out with one assume that the computing process is free of noises. In this paper, from a control-theoretical framework, noise-suppressing zeroing neural dynamic (NSZND) model is developed, analyzed and investigated by feat of continuous-time zeroing neural network model, which behaves efficiently for hurdling online time-varying nonlinear optimization problems with the presence of different noises. Further, for speeding the rate of convergence, general noise-suppressing zeroing neural network (GNSZNN) model with different activation functions is discussed. Then, theoretical analyses show that the proposed noise-suppressing zeroing neural network model derived from NSZND model has the global convergence property in the presence of different kinds of noises. Besides, how GNSZNN model performs with different activation functions is also proved in detail. In addition, numerical results are provided to substantiate the feasibility and superiority of GNSZNN model for online time-varying nonlinear optimization problems with inherent tolerance to noises.