No label defined (Q4258787)

Let \(G\) be a matrix finite dimensional Lie group and let \(g\) be its Lie algebra. Let \(S(G)\) be the class of stochastically continuous multiplicative semi-martingales taking values in \(G\), and let \(S(g)\) be the class of stochastically continuous additive semi-martingale taking values in \(g.\) The author proves that an arbitrary process \(X(t) \in S(G)\) is the exponent EXP of a process \(x(t) \in S(g)\) i.e. \[ X(t)= \lim_{|\pi |\to 0} \prod_{k=0}^{n_\pi -1} \exp\{ \Delta x(t_k^\pi)\}\colon = \exp(x)(t), \] in mean square, where \(\pi=\{ 0=t_0^\pi < t_1^\pi < \ldots t_n^\pi =t \}\), \(|\pi |=\max_k (t_k^\pi - t_{k-1}^\pi).\) It is known that an arbitrary process \(x(t) \in S(g)\) has the representation \(x(t) =x^c(t)+x^d(t)\), where \(x^c(t)\) is an additive continuous with probability one semi-martingale, \(x^d(t)\) is an additive discrete semi-martingale. The main result of this paper is the proof of the following representation of a process \(X(t)\) from \(S(G)\) \[ X(t)= \lim_{|\pi |\to 0} \prod_{k=0}^{n_\pi -1} \exp\{ \Delta x^c(t_k^\pi)\} \exp\{ \Delta x^d(t_k^\pi)\}, \] where \(x(t) =x^c(t)+x^d(t)\) is the process from the representation \(X(t)=\)EXP\((x)(t).\) For the entire collection see [Zbl 0896.00028].