zbMATH — the first resource for mathematics

On self attracting/repelling diffusions. (English) Zbl 1030.60075
A self-interacting diffusion is a continuous stochastic process which resolves the SDE of the form \[ dX_t= \sum_i F_i(X_t)\circ dB^i_t- {\alpha\over 2t} \Biggl(\int^t_0\nabla V_{X_s}(X_t) ds\Biggr) dt \] on a compact Riemannian manifold \(M\), where \((u,x)\to V_u(x)\) is a smooth function on \(M\times M\). The main purpose of this work is to study the empirical occupation measure \(\mu_t={1\over t} \int^t_0 \delta_{X_s}ds\). In the case where \(V\) has the form \[ V(x,y)= \int_C G(u,x) G(u,y)\nu(du)+ \beta, \] it is shown that the limit set \(L\{\mu_t\}\) of \(\{\mu_t; t> 0\}\) is a connected subset in the space of Borel probability measures \({\mathcal P}(M)\) over \(M\). Moreover any \(\mu\in L\{\mu_t\}\) admits a density which was characterized by an “energy function”. A sufficient condition on \(\alpha\) is given to insure that \(\mu^*= \lim_{t\to+\infty} \mu_t\) exists almost surely.

60J60 Diffusion processes
60F15 Strong limit theorems
Full Text: DOI arXiv
[1] Benaı̈m, M., Dynamics of stochastic approximation algorithms, (), 1-68 · Zbl 0955.62085
[2] Benaı̈m, M.; Ledoux, M.; Raimond, O., Self-interacting diffusions, Probab. theory related fields, 122, 1-41, (2002) · Zbl 1042.60060
[3] M. Benaı̈m, O. Raimond, Self-interacting diffusions III: The gradient case, in preparation
[4] Ben Arous, G.; Brunaud, M., Méthode de Laplace : étude variationnelle des fluctuations de type champ moyen, Stochastic, 31, 79-144, (1990) · Zbl 0705.60046
[5] Conley, C.C., Isolated invariant sets and the Morse index, CBMS regional conf. ser. in math., 38, (1978), American Mathematical Society Providence
[6] Tromba, A.J., The morse – sard – brown theorem for functionals and the problem of plateau, Amer. J. math., 99, 6, 1251-1256, (1977) · Zbl 0373.58003
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.