Some estimates of the transition density of a nondegenerate diffusion Markov process (Q1176149)

It is considered the difficult problem of estimating the transition density \(P_ t(x,y)\) for a nondegenerate diffusion process using Fleming's logarithmic transformation and the corresponding nonlinear dynamic programming equation of a stochastic control problem. Estimate of \(D^ m_ x \log P_ t(x,y)\) instead of \(D^ m_ xP_ t(x,y)\) brings difficulties which are avoided using a stochastic parallel translation. Two theorems and seven lemmas are proved.