Problem of eigenvalues of stochastic Hamiltonian systems with boundary conditions. (Q1877516): Difference between revisions
From MaRDI portal
Revision as of 19:39, 6 June 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Problem of eigenvalues of stochastic Hamiltonian systems with boundary conditions. |
scientific article |
Statements
Problem of eigenvalues of stochastic Hamiltonian systems with boundary conditions. (English)
0 references
7 September 2004
0 references
Let \(B\) be a standard one-dimensional Wiener process. A forward-backward stochastic differential equation \[ \begin{aligned} dx_ {t}&= \bigl [ H_ {21}x_ {t} + (1-\lambda )H_ {22} y_ {t} + (1-\lambda )H_ {23}z_ {t}\bigr ]dt+ \bigl [ (1-\lambda )H_ {31}x_ {t} + (1-\lambda )H_ {32}y_ {t} + H_ {33}z_ {t}\bigr ]dB_ {t},\\ -dy_ {t} &= \bigl [H_ {11}x_ {t} + H_ {12}y_ {t} + (1-\lambda )H_ {13}z_ {t}\bigr ]\,dt - z_ {t}\,dB_ {t}, \quad x(0)=0, \;y(T) = 0, \end{aligned} \tag{1} \] in \(\mathbb R^ {n}\) is considered. It is assumed that \(H_ {ij}\) are \((n\times n)\)-matrices such that \(H_ {ij} = H^ {T}_ {ji}\) and a monotonicity hypothesis \([\widehat H_ {ij}] \leq -\alpha I_ {3n}\) is satisfied for some constant \(\alpha >0\), where the \((3n\times 3n)\)-matrix \([\widehat H_ {ij}]\) is defined by \(\widehat H_ {1j} = - H_ {1j}\), \(\widehat H_ {ij} = H_ {ij}\) for \(i=2,3\). The system (1) may be viewed as a stochastic Hamiltonian system, the Hamiltonian function of which is a quadratic form. Obviously, \((x_ {t},y_ {t},z_ {t})\equiv 0\) is a trivial solution to (1); a real number \(\lambda \) is called an eigenvalue of the problem (1) if there exists a nontrivial solution (an eigenfunction) to (1). It is shown that there exists a smallest eigenvalue \(\lambda _ 1 >0\) of (1) and the dimension of the space of all eigenfunctions corresponding to \(\lambda _ 1\) is less than or equal to \(n\). If the space dimension \(n=1\), a more complete description of eigenvalues is available. For a system related to (1) it is proven that all eigenvalues form a countable set \(\lambda _ 1 <\lambda _ 2 <\dots \), \(\lambda _ {i}\uparrow +\infty \), and the dimension of the space of all eigenfunctions corresponding to any \(\lambda _ {i}\) is 1. Moreover it is shown that the problem may be reformulated as an eigenvalue problem for a suitable bounded self-adjoint operator in the Hilbert space \(M^ 2(0,T;\mathbb R^ {3n})\) of all \(\mathbb R^ {3n}\)-valued adapted stochastic processes that are square integrable over \(\Omega \times (0,T)\).
0 references
stochastic Hamiltonian systems
0 references
forward-backward stochastic differential equations
0 references
matrix-valued Riccati equations
0 references
0 references
0 references
0 references