Positive polynomials and semidefinite programming (Q948606)

This is a short survey (in German) of relatively recent (1998--2008) research activities at the border between optimization, numerical linear algebra, functional analysis and real algebraic geometry. The focus in on the use of semidefinite programming (linear programming in the convex cone of symmetric matrices with non-negative eigenvalues, see Section 5) to solve non-convex polynomial optimization problems (Section 6) or certify emptiness of semi-algebraic sets. These techniques rely on the primal problem of moments (Section 4) consisting in finding conditions under which an infinite or truncated sequence of real numbers represent moments of a measure supported on a given semi-algebraic set, and the dual problem of representing a polynomial positive on this set (Section 3). The survey concludes by sketching some current research directions in Section 7. For comprehensive surveys (in English) covering this material and much more, this reviewer recommends [\textit{J. W. Helton} and \textit{M. Putinar}, ``Positive polynomials in scalar and matrix variables, the spectral theorem and optimization'', in: Operator theory, structured matrices, and dilations, Theta, Bucharest, 229--306 (2007; Zbl 1199.47001)] and [\textit{M. Laurent}, in: Emerging applications of algebraic geometry, Vol. 149 of IMA Volumes in Mathematics and its Applications, Springer, 157--270 (2009; Zbl 1163.13021)].