Laws of the iterated logarithm for some symmetric diffusion processes (Q1591240)

The authors prove three results concerning laws of iterated logarithms for certain functionals of some Markov process. First a large deviation principle similar to that of Schilder for Brownian motion is proved for some symmetric diffusions as Brownian motion on Sierpiński gaskets or on Sierpiński carpets (in fact for diffusion whose densities satisfy inequalities of Aronson's type) [see also \textit{M. T. Barlow} and \textit{R. F. Bass}, Can. J. Math. 51, No. 4, 673-744 (1999; Zbl 0945.60071)] and \textit{P. J. Fitzsimmons}, \textit{B. M. Hambly} and \textit{T. Kumagai}, Commun. Math. Phys. 165, No. 3, 595-620 (1994; Zbl 0853.60062)]. The result is then applied to obtain the law of the iterated logarithm. Then the case of arbitrary Markov processes not necessarily continuous nor symmetric is considered. An upper bound for the law of the iterated logarithm is given for functionals that are nondecreasing continuous, subadditive satisfying a uniform scaling property. Finally, the law of the iterated logarithm is studied for the supremum in \(x\) of \(\ell_{t}(x)\), the local time of the Brownian motion whose state space is an affine nested fractal or a Sierpiński carpet. Also the law of the iterated logarithm for the \(\mu\)-range of such processes is studied (\(\mu\) is an invariant measure for the state space).