arXiv subjects
Publications and source records attributed to G. B. Levitina.
We prove that every derivation acting on a von Neumann algebra $\mathcal{M}$ with values in a quasi-normed bimodule of locally measurable operators affiliated with $\mathcal{M}$ is necessarily inner.
Let $\mathcal{A}$ be a commutative $AW^*$-algebra, let $S(\mathcal{A})$ be the *-algebra of all measurable operators affiliated with $\mathcal{A}$, let $\mathcal{I}$ be an ideal in $\mathcal{A}$, let $s(\mathcal{I})$ be the support of the ideal $\mathcal{I}$ and let $\mathbb{Y}$ be a solid subspace in $S(\mathcal{A})$. The necessary and sufficient conditions of existence of non-zero band preserving derivations from $\mathcal{I}$ to $\mathbb{Y}$ are given. We show that, in case when $\mathbb{Y}\subset\mathcal{A}$, or $\mathbb{Y}$ is a quasi-normed solid space, any band preserving derivation from $\mathcal{I}$ into $\mathbb{Y}$ is always trivial. At the same time, there exist non-zero band preserving derivations from $\mathcal{I}$ with values in $S(\mathcal{A})$, if and only if the Boolean algebra of all projections from the $AW^*$-algebra $s(\mathcal{I})\mathcal{A}$ is not $σ$-distributive.
Let $\mathcal{I,J}$ be symmetric quasi-Banach ideals of compact operators on an infinite-dimensional complex Hilbert space $H$, let $\mathcal{J:I}$ be a space of multipliers from $\mathcal{I}$ to $\mathcal{J}$. Obviously, ideals $\mathcal{I}$ and $\mathcal{J}$ are quasi-Banach algebras and it is clear that ideal $\mathcal{J}$ is a bimodule for $\mathcal{I}$. We study the set of all derivations from $\mathcal{I}$ into $\mathcal{J}$. We show that any such derivation is automatically continuous and there exists an operator $a\in\mathcal{J:I}$ such that $δ(\cdot)=[a,\cdot]$, moreover $\|a\|_{\mathcal{B}(H)}\leq\|δ\|_\mathcal{I\to J}\leq 2C\|a\|_\mathcal{J:I}$, where $C$ is the modulus of concavity of the quasi-norm $\|\cdot\|_\mathcal{J}$. In the special case, when $\mathcal{I=J=K}(H)$ is a symmetric Banach ideal of compact operators on $H$ our result yields the classical fact that any derivation $δ$ on $\mathcal{K}(H)$ may be written as $δ(\cdot)=[a,\cdot]$, where $a$ is some bounded operator on $H$ and $\|a\|_{\mathcal{B}(H)}\leq\|δ\|_\mathcal{I\to I}\leq 2\|a\|_{\mathcal{B}(H)}$.