Itô formula for free stochastic integrals (Q1604551): Difference between revisions

The paper treats the stochastic integration in the context of free probability. Based on the Itô isometry, the author defines stochastic integrals \(\int_0^{\infty} U(s) \# \text{d} X(s)\) of adapted biprocesses \(U\) with respect to adapted free stochastic measures \(X(t)=X( [0,t))\) with distribution \(\mu\), and investigates their properties, where \(\#\) is the left bimodule action : \(( {\mathcal A} \otimes {\mathcal A}^{op}) \times {\mathcal A}\rightarrow {\mathcal A}\) and \({\mathcal A}^{op}\) is the opposite algebra of the basic von Neumann algebra \({\mathcal A}\) in \(W^*\)-noncommutative probability space \(( {\mathcal A}, \varphi)\). Such integrals for square-integrable integrands in the appropriate sense as well as for a certain general class of bounded integrands are considered. Using the product form of the Itô formula, the author proves the full functional Itô formula in the framework of free probability. The author restricts the analysis to the free stochastic measures consisting of bounded operators, which provides convergence in the operation norm (meaning a strong sense) both for the integrals and for the limits defining the higher diagonal measures. This is one of the peculiar features of his work. In fact, the lower bound in the free Burkholder-Davis-Gundy inequality is derived, which enables the author to prove the extended Itô product formula and the extended functional Itô formula for the wider class of bounded biprocesses. The paper is based on author's previous work [Adv.\ Math. 155, No.~1, 154-179 (2000; Zbl 0978.46041)], and also due to \textit{P. Biane} and \textit{R. Speicher} [Probab.\ Theory Related Fields 112, No.~3, 373-409 (1998; Zbl 0919.60056)] \textit{G.-C. Rota} and \textit{T. C. Wallstrom} [Ann.\ Probab. 25, No.~3, 1257-1283 (1997; Zbl 0886.60046)].