Dynamics of almost periodic scalar parabolic equations (Q1900951): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Set OpenAlex properties. |
||
(One intermediate revision by one other user not shown) | |||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank | |||
Property / OpenAlex ID | |||
Property / OpenAlex ID: W2094247340 / rank | |||
Normal rank |
Revision as of 21:49, 19 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Dynamics of almost periodic scalar parabolic equations |
scientific article |
Statements
Dynamics of almost periodic scalar parabolic equations (English)
0 references
23 June 1996
0 references
The paper is devoted to the scalar parabolic equation \[ u_t=u_{xx}+ f(t, x, u, u_x),\quad t> 0,\quad 0< x< 1,\tag{1} \] with the boundary conditions \[ \beta u(t, 0)+ (1- \beta) u_x(t, 0)=0,\quad \beta u(t, 1)+ (1- \beta) u_x(t, 1)=0,\;t> 0,\tag{2} \] where \(\beta=0\) or 1, \(f: R^1\times [0, 1]\times R^1\times R^1\to R^1\) is \(C^2\), and \(f(t, x, u, p)\) with all its partial derivatives up to order 2 is almost periodic in \(t\) uniformly for \((x, u, p)\) in compact sets. A dynamical system is associated to (1)--(2) in the following way. Let \(C=C(R^1\times [0, 1]\times R^1\times R^1, R^1)\) be the space of continuous functions \(F: R^1\times [0, 1]\times R^1\times R^1\to R^1\) with the topology of the uniform convergence on compact subsets. It follows from classical topological dynamical system theory that the time translation \((F, t)\to F_t: F_t(s, x, u, p)=F(t+ s, x, u, p)\) defines a flow on \(C\), and the hull \(H(f)=\text{cl}\{f_t\mid t\in R^1\}\) of \(f\) is an almost periodic minimal set, that is \(H(f)\) is minimal and each motion in \(H(f)\) is almost periodic. Each \(g\in H(f)\) is also a \(C^2\) function and \(H(f)\) gives rise to a family of equations associated to each \(g\in H(f)\), \[ u_t=u_{xx}+ g(t, x, u, u_x),\quad t> 0,\quad 0< x< 1,\tag{3} \] subject to the boundary conditions (2). Let \(X\) be a fractional power space associated with the operator \(u\to - u_{xx}: {\mathcal D}\to L^2(0, 1)\) being embedded in \(C^1[0, 1]\), where \({\mathcal D}=\{u\mid u\in H^2(0, 1)\), \(u\) satisfies (2)\}. Then \(F: R^1\times X\to L^2(0, 1)\), \(F(t, u)(x)=f(t, x, u, u_x)\), is well defined, and for any \(U\in X\) problem (3)--(2) admits (locally) a unique solution \(u(t,\cdot, U, g)=U(\cdot)\). This solution depends continuously on \(g\in H(f)\) and \(U\in X\) and defines a semiflow \(\Pi_t\) on \(X\times H(f)\), \[ \Pi_t(U, g)=(u(t, \cdot, U, g), g\cdot t),\quad t> 0,\tag{4} \] where \(g\cdot t\) is the flow on \(H(f)\) defined by the time translations. In the terminology of the semiflow (4) the study of asymptotic behavior for a bounded solution \(u(t, x)\) of (1)--(2) gives rise to the problem of understanding the \(\omega\)-limit set \(\omega(U_0, f)\) of the bounded motion \(\Pi_t(U_0, f)\) in \(X\times H(f)\), where \(U_0(x)=u(0, x)\) (it is known that if \(u(t, \cdot, U, g)\) is bounded in \(X\) for \(t\) in the existence interval of the solution, then \(u\) is globally defined classical solution). In the case that \(f\) is time periodic with period \(T\), it is well-known that each bounded solution \(u(t, x, U_0, f)\) of (1)--(2) approaches a periodic solution with period \(T\). In the language of the semiflow (4) this means that each \(\omega\)-limit set \(\omega(U,g)\) \((g\in H(f))\) is a periodic minimal set in \(X\times H(f)\) with period \(T\). Similar results are false in general for time almost periodic problems (1)--(2). There are examples in scalar ODE, which show that the \(\omega\)-limit sets of (4) may not be minimal, or they may not be almost periodic minimal even if minimal. Two natural questions arise in the study of (1)--(2): (1) What kind of structure one can expect for an \(\omega\)-limit set \(\omega(U, g)\) of (4) if it is not minimal? (2) Does an \(\omega\)-limit set still carry over some ``oscillation'' properties of the original system (1)--(2) if it is not an almost periodic minimal set? This paper gives partial answers to these questions. It is proved that for the semiflow (4), each \(\omega\)-limit set \(\omega(U, g)\) contains at most two (obviously at least one) minimal invariant sets, and each minimal invariant set contained in \(\omega(U, g)\) is a proximal extension of \(H(f)\). In the case, where two minimal invariant sets appear in the \(\omega\)-limit set \(\omega(U, g)\), both are almost automorphic extensions of \(H(f)\). If \(\omega(U, g)\) is distal or almost periodic minimal, then it must be an almost periodic extension of \(H(f)\). There is an example showing that an \(\omega\)-limit sets of (4) may contain two minimal sets, and at least one of them is an almost automorphic but not an almost periodic extension of \(H(f)\).
0 references
\(\omega\)-limit set
0 references
almost periodic minimal set
0 references