Integrals of operator-valued functions (Q1101918)

\textit{J. Mikusiński} [The Bochner integral (1978; Zbl 0369.28010)] proposed a very simple scheme to define the Lebesgue and Bochner integrals. In the case of the Lebesgue integral, in one dimension, this scheme is as it follows: Let \(\{f_ n\}\) be a sequence of the characteristic functions of a sequence of finite intervals and \(\lambda_ n\) real numbers. Suppose that the numerical series (1) \(\sum_{n\geq 0}| \lambda_ n| \int_{R}f_ n(x)dx\) converges and the functional series (2) \(\sum_{n\geq 0}\lambda_ nf_ n(x)\) converges absolutely in some points \(x\in A\), then series (2) defines a function f in the points \(x\in A\) and by definition \(\int f(x)dx=\sum_{n\geq 0}\lambda_ n\int_{R}f_ n(x)dx.\) It is easy to enlarge this scheme to the n-dimensional case to the Bochner integral. The author of this paper uses the Mikusiński scheme to define an integral of a distribution valued function or Mikusiński operator valued function. He takes instead of the series (1) and (2) the following ones: In the case of a distribution valued function the series \(\sum_{n\geq 0}| <\lambda_ n,\phi >| \int_{R}f_ n(x)dx\) and \(\sum_{n\geq 0}<\lambda_ n,\phi >f_ n(x)dx\), where \(\lambda_ n\in D'\) and \(\phi\in D\). In the case of a Mikusinski operator valued function \(\sum_{n\geq 0}\| \sigma \lambda_ n\|_ m\int_{R}f_ n(x)dx\) and \(\sum_{n\geq 0}(\sigma \lambda_ n)(t)f_ n(x)\), where \(\lambda_ n\) belongs to the Mikusinski field \({\mathcal M}\), \(\sigma\in C[0,\infty)\) and such that \(\sigma \lambda_ n\in C[0,\infty)\), \(n\in N\); \(\| g\|_ m=\max | g(x)|,\) \(x\in [0,\infty).\) In both cases he introduces an expansion \(f\cong \sum \lambda_ nf_ n\). The paper analyses different situations in the second case enlarging this expansion to the Mikusiński field of operators in two variables. It is more an introduction in a new approach to the integrals in the generalized function spaces. We can only expect that the author will continue with this research.