Linear systems in a saturated mode and convergence as gain becomes large of asymptotically stable equilibrium points of neural nets (Q1971585)

The paper investigates the existence and regularity properties of the solutions of system (M): \( x' \in Tx+c- \partial I_{D^{n}}x\), called ``linear system in saturated mode''. In the paper of \textit{J. H. Li, A. N. Michel} and \textit{W. Porod} [IEEE Trans. Circuits Systems 36, No. 11, 1405-1442 (1998; Zbl 0689.94004)] it was conjectured that there was a one-to-one correspondence between the sets of asymptotically stable equilibrium points for the system (M) and the system (H): \( Cu'=Tv+c-R^{-1}u\), \(v=G(\lambda u)\). The author shows that this conjecture is correct. The conditions required for it to hold are also analysed.