Some asymptotic estimates of transition probability densities for generalized diffusion processes with self-similar speed measures (Q1173970)

This paper deals with strong Markov processes associated with generalized diffusion operators \(A={d\over dm(x)}{d\over dx}\) in the case where the measure \(m\) is self-similar; self-similarity of sets and measures has been defined by \textit{J. E. Hutchinson} [Indiana Univ. Math. J. 30, 713- 747 (1981; Zbl 0598.28011)] through a family of contraction affine maps and this has given a rigorous setting to the fractal theory. If \(p(t,x,y)\) is the transition probability density of \(X\) with respect to \(dm\), the main result is the equivalence between \(-\log p(t,x,y)\) and \(t^{-s}\) as \(t\) goes to 0, where \(s\) is the similarity dimension of the measure \(m\). More precise estimates are given under additional hypotheses and some relations between the similarity dimension, the entropy and the Kolmogorov dimension are investigated.