scientific article; zbMATH DE number 4090521

From MaRDI portal
Publication:3817403





Let (X,\({\mathcal B},\mu)\) be a metric measure space, \(\{x_ n\}_ 0^{\infty}\) a Markov process with state space X and one-step transition probability density \(q_{\sigma}(x,y)\), where \(\sigma >0\) is a parameter. Next, let f be a nice map from X into itself. The stochastic perturbation system \(f_{\sigma}\) considered in the paper evolves as follows: At first the system is moving from \(x\in X\) to \(x_{\sigma}\) according to the Markovian rule, but it chooses the real position f \(x_{\sigma}=f_{\sigma}\) \(x\in X\) under the control of f. Set \[ S_{\sigma}(\underline x,n,N;\epsilon)={\mathbb{P}}[f^ m_{\sigma}x_ n\in B_{\epsilon}(x_{n+m}),\quad 1\leq n\leq N] \mu (B_{\epsilon}(0))^{-N} \] where \b{x}\(=(x_ 1,x_ 2,...)\), \(n,N\in {\mathbb{Z}}_+\), \(B_{\epsilon}(x)\) is the ball with center x and radius \(\epsilon\) and \(f^ m_{\sigma}\) is the m th composition of \(f_{\sigma}\). The author is interested in studying the deterministic quantitiy of the system: \[ S_{\sigma}(\underline x,n,N)=\lim_{\epsilon \to \infty}\log S_{\sigma}(\underline x,n,N;\epsilon). \] This is the so-called Onsager-Machlup functional of the system. Under some assumptions on \(q_{\sigma}\) and f, the author obtains an explicit expression and some asymptotic expansions of \(S_{\sigma}\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3817403)