Convex cones in mapping spaces between matrix algebras (Q2226432)

Let \(\mathbb{CP}\) denote the convex cone of all completely positive maps in a matrix algebra. A closed convex one-sided cone \(K\) of positive linear maps satisfies \(K \circ \mathbb{CP} \subset K\) (or \(\mathbb{CP} \circ K \subset K\)). This is a relaxation of the notion of (two-sided) cones introduced in [\textit{E. Størmer}, J. Funct. Anal. 66, 235--254 (1986; Zbl 0637.46061)] and the paper extends many properties of two-sided cones to one-sided ones. Furthermore, one-sided cones admit a characterisation in terms of ampliation \(\phi\mapsto 1 \otimes \phi\) or \(\mapsto \phi \otimes 1\) of a linear map \(\phi\). The obtained results can be used in quantum information theory. Several equivalent statements of the PPT (positive partial transpose) square conjecture are obtained as well.