Metastability in reversible diffusion processes. I: Sharp asymptotics for capacities and exit times (Q1766229)

The authors investigate reversible diffusion processes \(X_{\varepsilon}(t)\), governed by an Itô stochastic differential equation \(dX_{\varepsilon}(t)=-\nabla F(X_{\varepsilon}(t))\,dt+ \sqrt{2\varepsilon}\,dW(t)\) on a regular domain \(\Omega\subseteq \mathbb R^d,\) where the drift \(\nabla F\) is generated by a sufficiently regular potential function. It is supposed that \(X_{\varepsilon}\) is vanished on \(\Omega^{c}\) and the function \(F(x)\) has several local minima. It is shown that metastable exit times from the attractive domains of the minima \(F\) can be related, up to multiplicative errors tending to one as \(\varepsilon\downarrow 0\), to the capacities of suitably constructed sets. In terms of local characteristies of \(F\) at the starting minimum and the relevant saddle points these capacities can be computed up to multiplicative errors tending to one. On the base of these results the first rigorous proof of the classical Eyring-Kramers formula in dimension larger than 1 is given. Part II, cf. Metastability in reversible diffusion processes. II: Precise asymptotics for small eigenvalues. ibid. 7, No. 1, 69--99 (2005; Zbl 1105.82025).