Stochastic integration of operator-valued functions with respect to Banach space-valued Brownian motion (Q934818)

Let \(E\) be a Banach space with property \((\alpha)\) and let \(W_\Gamma\) be an \(E\)-valued Brownian motion with distribution \(\Gamma\). Property \((\alpha)\) means that, given mutually independent Rademacher sequences \((r'_n)_{n\geq 1}\), \((r''_n)_{n\geq 1}\), on probability spaces \((\Omega',{\mathbf P}')\), \((\Omega'',{\mathbf P}'')\), there exists a constant \(C>0\) such that \[ {\mathbf E}'{\mathbf E}''\left\| \sum_{m,n=1}^N \varepsilon_{mn} r'_m r''_n x_{mn}\right\|^2 \leq C\, {\mathbf E}''{\mathbf E}'\left\| \sum_{m,n=1}^N r'_m r''_n x_{mn}\right\|^2 \] for all \(N\geq 1\) and all choices \(\varepsilon_{mn}\in\{0,1\}\) and \(x_{mn}\in E\). This property was introduced by \textit{G. Pisier} [Ann. Math. (2) 115, 375--392 (1982; Zbl 0487.46008)] who proved that every Banach space with local unconditional structure and finite cotype has the property \((\alpha)\). Replacing Rademacher sequences, by orthonormal Gaussian sequences the authors define the Gaussian property \((\alpha)\) and prove that it is equivalent to the property \((\alpha)\). Exploiting this property, they show that a function \(\Psi\) on \([0,T]\) with values in bounded linear operators on \(E\) is stochastically integrable with respect to \(W_\Gamma\) if and only if \(\Gamma\)-almost all orbits \(\Psi x\) are stochastically integrable with respect to a real Brownian motion. This result is derived from an abstract result on the existence of \(\Gamma\)-measurable linear extensions of \(\gamma\)-radonifying operators (a bounded operator \(T\) on a Hilbert space \(H\) to \(E\) is \(\gamma\)-radonifying if, for some orthonormal basis \((h_n)_{n\geq 1}\) of \(H\) and a Gaussian series \((g_n)_{n\geq 1}\), the Gaussian series \(\sum_{n\geq 1}g_nTh_n\) converges in \(L^2(\Omega;E)\)). As an application, they obtain a necessary and sufficient condition for solvability of stochastic evolution equations driven by an \(E\)-valued Brownian motion.