A family of varying-parameter finite-time zeroing neural networks (VPFTZNN) is introduced for solving the time-varying Sylvester equation (TVSE). The convergence speed of the proposed VPFTZNN family is analysed and compared with the traditional zeroing neural network (ZNN) and the finite-time zeroing neural network (FTZNN). The behaviour of the proposed neural networks under various activation functions is proved theoretically and verified experimentally. In addition, the stability and noise resistance of the proposed VPFTZNN family are discussed. Further, the proposed VPFTZNN models are applied in the computation of current flows in an electrical network. (c) 2021 Elsevier B.V. All rights reserved.
NSTL
Elsevier