A stability theory for integral equations (Q1899922)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A stability theory for integral equations
scientific article

    Statements

    A stability theory for integral equations (English)
    0 references
    18 March 1996
    0 references
    The most important thing we have to know in stability theory of differential equations is the set of what we call equilibrium points. A useful meaning in developing such a theory is introduced in the present paper and it is termed ``near equilibrium''. Briefly speaking given an integral equation \(x(t)=a(t)-\int^t_{\alpha(t)} Q (t, s, x(s))ds\) we say that a pair of functions \((\psi, \Psi)\) is a near equilibrium if \(\Psi(t) := a(t)-\psi(t)-\int^t_{\alpha (t)} Q(t,s, \psi (s))ds \in {\mathcal L}^1 [0, +\infty)\). In other words \(\psi\) fails to be a solution by an amount of an \({\mathcal L}^1\) function. The authors develop a theory concerning asymptotic stability (in a certain sense) of a near equilibrium of the above integral equation. The proofs of their stability results rely mainly on the use of Lyapunov-like functions. A characteristic result of the paper is the following: Theorem 1B: Consider the equation \(x(t)=a(t)-\int^t_{t-h} D(t,s) x(s) ds\), where assume that \(a : \mathbb{R} \to \mathbb{R}\) is continuous, \(a\) and \(a^2 \in {\mathcal L}^1[0,\infty)\), \(D(t,s) \geq 0\), \(D_s(t,s) \geq 0\), \(D_{\text{st}} (t,s) \leq 0\), \(D(t,t-h)=0\) and \(D(t,t) \leq P\) for some \(P > 0\). Then there exist continuous functions \(p, q : [0, \infty) \to [0,\infty)\), \(p(t) \to 0\) as \(t \to\infty\), \(q \in {\mathcal L}^1 [0, \infty)\), a continuous function \(V(t,x(.))\) defined for a solution \(x(t)=x(t,0,\varphi)\), where \(\varphi\) satisfies \(\varphi(0)=a(0)-\int^0_{-h} D(0,s) \varphi(s) ds\), and wedges \(W_i\), \(i=1,2,3\), such that \(W_1(r) \to \infty\) as \(r \to \infty\), \(W_1(|x(t)-a(t)|) \leq PV(t,x(.)) \leq W_2(|||(x-a)_t |||) + p(t)\) and \(V'(t,x (.)) \leq -W_3 (|x(t)-a(t)|) + q(t)\). (Notice that the symbol \(|||\cdot |||\) stands for the \(L^2\)-norm and as usually \(q_t(s)=q(t + s)\), \(s\in[-h, 0]\).) Several examples illustrate the results of the paper in case of finite or infinite delay, where the stability regions are also provided.
    0 references
    0 references
    0 references
    0 references
    0 references
    near equilibrium
    0 references
    stability theory
    0 references
    equilibrium points
    0 references
    asymptotic stability
    0 references
    integral equation
    0 references
    Lyapunov-like functions
    0 references
    infinite delay
    0 references
    stability regions
    0 references
    0 references
    0 references
    0 references