Linear optimization with cones of moments and nonnegative polynomials (Q745685)

This paper studies properties of the cone of polynomials and its dual cone by characterizing their interiors, proving their closeness and dual relationship. A convergent hierarchy of semidefinite relaxations for each of them is constructed. A semidefinte algorithm is proposed to solve linear optimization problems with the two cones. The asymptotic and finite convergence of the algorithm is established. The author shows that if the optimization problems are feasible then a feasible point is obtained, otherwise a certificate for the infeasibility is provided.