On the \(R\)-boundedness of stochastic convolution operators (Q2355144)

The paper presents a systematic study of the so called \(R\)-boundedness of sets of vector-valued stochastic convolution operators with scalar-valued kernels. This topic is a prerequisite to the theory of stochastic maximal \(L^p\)-regularity for vector-valued stochastic convolutions in the following sense: in order to say whether a stochastic convolution process satisfies a maximal \(L^p\)-regularity estimate, one needs to know whether a certain class of vector-valued stochastic convolution operators with scalar-valued kernels is \(R\)-bounded. In this direction, \(R\)-boundedness of sets of stochastic convolution operators is related to \(\ell^s\)-boundedness of the corresponding sets of deterministic convolution operators, to the UMD property, the type and the martingale type of the state space \(X\), and to the Hardy-Littlewood property of the dual \(X^*\) and the dual of the \(2\)-concavification \((X^2)^*\). To summarize, the authors prove sufficient and necessary conditions for \(\ell^s\)-boundedness of sets of deterministic convolution operators, sufficient conditions for \(R\)-boundedness of sets of stochastic convolution operators and an equivalence between \(R\)-boundedness of sets of stochastic convolution operators, \(R\)-boundedness of the corresponding sets of deterministic convolution operators and \(\ell^1\)-boundedness of ``mirrored'' sets of deterministic convolution operators. Moreover, a set of stochastic convolution operators is \(R\)-bounded if and only if the set of the restrictions of these operators to deterministic integrands is \(R\)-bounded. Finally, the authors complement the positive results by a counterexample showing a failure of \(R\)-boundedness of a rather natural set of stochastic convolution operators on the Banach lattice \(\ell^2(\ell^4)\).