Noncommutative Positivstellensätze for pairs representation-vector (Q655985)

Positivity in non-commutative *-algebras was recently resurrected by several groups of researchers for specific theoretical or applied reasons. We cite only the recent works [\textit{K. Schmüdgen}, Math. Ann. 331, No. 4, 779--794 (2005; Zbl 1137.47030); Math. Z. 254, No. 3, 641--653 (2006), erratum 259, No. 1, 231 (2008; Zbl 1131.14061); \textit{M. C. de Oliveira, J. W. Helton, S. A. McCullough} and \textit{M. Putinar}, in: The IMA Volumes in Mathematics and its Applications 149, 17--61 (2009; Zbl 1156.14331); \textit{S. Pironio, M. Navascués} and \textit{A. Acín}, SIAM J. Optim. 20, No. 5, 2157--2180 (2010; Zbl 1228.90073); \textit{M. Brešar} and \textit{I. Klep}, Math. Res. Lett. 16, No. 4, 605--626 (2009; Zbl 1189.16021)]. In the context of such an explosion of free-* Positivstellensätze, the present note by Cimpric offers a lucid and final algebraic explanation of the major technical aspects of this new field of research. The author proposes a duality between convex cones in a *-algebra \(A\) and positivity sets consisting of pairs \((\pi,v)\), where \(\pi\) is an unbounded *-representation of \(A\) and \(v\) is a selected vector in the representation space. The proofs of the major results are optimal, and have the flavor of reaching their final form. These will be the framework and proofs I will use in any future lectures devoted to positivity in a *-algebra.