On p-convexity and q-concavity of unitary matrix spaces (Q1062240): Difference between revisions

Let E be a symmetric sequence space. The associated unitary matrix space \(C_ E\) consists of all operators x on \(\ell_ 2\) so that the sequence \(s(x)=(s_ n(x))^{\infty}_{n=1}\) of singular numbers (i.e., the eigenvalues of \(| x| =(x^*x)^{1/2})\) belongs to E, with the norm \(\| x\|_{C_ E}=\| s(x)\|_ E\). If \(\{e_ n\}\), \(\{f_ n\}\) are orthonormal sequences in \(\ell_ 2\), then the map \(E\ni (a_ n)^{\infty}_{n=1}\to^{V}\sum^{\infty}_{n=1}a_ n(\cdot,e_ n)f_ n\in C_ E\) is an isometry and there is a contractive projection from \(C_ E\) onto V(E). Therefore many geometrical properties of E extend to \(C_ E.\) In the present paper the authors study two such properties. One says that \(C_ E\) is p-convex (or, p-concave) if there exists \(K>0\) so that for every finite sequence \(x_ 1,...,x_ n\) in \(C_ E\) holds \[ \| (\sum^{n}_{j=1}| x_ j|^ p)^{1/p}\|_{C_ E}\leq K(\sum^{n}_{j=1}\| x_ j\|^ p_{C_ E})^{1/p} \] \[ (respectively,\quad \| (\sum^{n}_{j=1}| x_ j|^ p)^{1/p}\|_{C_ E}\geq K^{-1}(\sum^{n}_{j=1}\| x_ j\|^ p_{C_ E})^{1/p}). \] Theorem \(A: C_ E\) is p-convex (or, p-concave) if and only if E is p-convex (respectively, \(p\)-concave). C\({}_ E\) is said to have an upper-p-estimate if there exists a constant M so that \[ \| \sum^{n}_{j=1}x_ j\|_{C_ E}\leq M(\sum^{n}_{j=1}\| x_ j\|^ p_{C_ E})^{1/p} \] for all finite sequences \(\{x_ j\}^ n_{j=1}\) of right-disjointly supported elements of \(C_ E\) (i.e. \(x^*_ ix_ j=0\) if \(i\neq j)\). The notion of lower-p-estimate in \(C_ E\) is defined in the obvious way. Theorem B: If E if p-convex (or, q-concave) then \(C_ E\) has an upper-r- estimate (respectively, lower-s-estimate) where \(r=\min \{p,2\}\) (respectively, \(s=\max \{q,2\})\).