Global solvability of time-varying semilinear differential-algebraic equations, boundedness and stability of their solutions. II (Q2659262)

This paper is the second part of [Differ. Equ. 57, No. 1, 19--40 (2021; Zbl 1466.34014); translation from Differ. Uravn. 57, No. 1, 22--42 (2021)] by the same author, it contains the Lyapunov stability and a detailed example. The semilinear DAEs considered read \[\frac{d}{dt}[A(t)x]+B(t)x = f(t,x),\] for \(t\geq t_+, \: t_0\geq t_+\geq 0\), \(x(t_0)=x_0\) be given. The Lyapunov stability is studied around the equilibrium solution \(x_\star(t)\equiv 0\) of the problem, when \(f(t,0)\equiv 0\). The stability and asymptotical stability, for the respective forms \(\frac{d}{dt}[Ax]+Bx = f\) and \(A\frac{dx}{dt}+Bx = f\) of the DAEs, are exposed in Theorems 3.1 and 3.4 respectively. Their proofs rely on hypotheses on the associated pencil \(\lambda A(t) + B(t)\), the derivatives of the map \(f\), the consistency manifold and the existence of a convenient Lyapunov map \(V(t,x)\). Stronger conditions are introduced by the authors in Theorems 3.3 and 3.6, to ensure the global asymptotic stability of the equilibrium point \(x_\star\). A useful section explains the efficient choice of the Lyapunov map, typically a bounded self-adjoint time-dependent operator \(t\mapsto H(t)\) to create a map \[V(t,x(t)) = (x(t), H(t)x(t)).\] The example is an electrical network, containing variable resistances, conductances and inductances; it is modelled by a 3-dimensional semilinear DAEs, the unknowns being the intensities. The \(3\times3\)-matrix \(A(t)\) is of rank \(1\) which induces two algebraic constraints. The resolution and the application of stability conditions is exposed well and allows to understand the theoretical content of both parts of this work. The stability theorems should be exploited to automate the mixed symbolic and numerical study of such systems in low dimension, based on linear algebra.