Shilnikov saddle-focus homoclinic orbits from numerics: higher dimensions (Q2116437)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Shilnikov saddle-focus homoclinic orbits from numerics: higher dimensions |
scientific article |
Statements
Shilnikov saddle-focus homoclinic orbits from numerics: higher dimensions (English)
0 references
17 March 2022
0 references
The authors give computable criteria for verifying three conditions: general position, asymptotic tangency and a transversality condition. They consider the differential equation \(\dot{x}=f(x)\), \(x\in \mathbb{R}^{n}\), where \(f\) is \(C^{r}\) (\(r\geq 1\)). Further, they suppose that \(z\) is a hyperbolic equilibrium and that \(q(t)\) is an associated homoclinic orbit. Furthermore, they call \(q(t)\) a Shilnikov saddle-focus homoclinic orbit if the following four conditions are satisfied: (D1) The eigenvalues of \(f^{\prime}(z)\) having the (strictly) smallest positive real part are \(\beta\pm i\omega\) with \(\omega>0\), each having algebraic multiplicity one, and satisfying \(0<\beta<-\mathrm{Re}(\lambda)\) for all eigenvalues \(\lambda\) with negative real parts; (D2) Up to a scalar multiple, \(\dot{q}(t)\) is the only bounded solution of \(\dot{y}=f^{\prime}(q(t))y\); (D3) \(\dot{q}(t) e^{-\nu t}\) is not bounded on \(\mathbb{R}_{-}\), where \(\beta<\nu<\mathrm{Re}(\lambda)\) for all eigenvalues \(\lambda\) with \(\mathrm{Re}(\lambda)>\beta\); (D4) It does not exist \(\xi\neq 0\) such that the solution \(y(t)\) of \(\dot{y} = -f^{\prime}(q(t))^{T} y\), \(y(0) = \xi\) is bounded on \(\mathbb{R}_{-}\) and the solution \(y(t)\) of \(\dot{y} = -[f^{\prime}(q(t))^{T}-\nu] y\), \(y(0) = \xi\) is bounded on \(\mathbb{R}_{+}\). The authors explain how one can varify that this orbit satisfies the four conditions (D1--D4). It turns out that the invertibility of a certain matrix \(L\) together with additional smallness conditions imply (D2). Conditions (D3) and (D4) are implied by the existence of left inverses for two other matrices together with additional smallness conditions. The authors study the rigorous numerical verification in four dimensions of the existence of a homoclinic orbit in parametrized autonomous systems when an approximate homoclinic orbit is given. The following example of a 4-dimensional system depending on two parameters \(\kappa\) and \(\gamma\) is considered: \(\dot{x}_{1}=x_{2}\), \(\dot{x}_{2}=-x_{1}+x_{2}+\kappa x_{3}\), \(\dot{x}_{3}=x_{4}\), \(\dot{x}_{4}=-g(x_{3})+x_{1}+\gamma x_{4}\), with \(g(y)=2y^{3}-y\). This system is obtained by perturbing a system satisfying conditions (D1), (D2) and (D4). The authors show graphically some numerically computed approximate homoclinic orbits for different parameters \(\kappa\) and \(\gamma\). Moreover, they give examples for which all conditions (D1--D4) or only some of them are not satisfied. The paper is supported by many numerical experiments.
0 references
Shilnikov chaos
0 references
homoclinic
0 references
shadowing
0 references
rigorous numerics
0 references
0 references