J. Vicente Pérez, V. Jeyakumar
In this talk we examine necessary and sufficient conditions for exact stable relaxations of a class of convex polynomial optimization problems in the face of data uncertainty in the constraints. We consider a robust SOS-convex polynomial optimization problem where the constraint data is affinely parameterized and the uncertainty sets are assumed to be bounded spectrahedra. The class of SOS-convex polynomials is a numerically tractable subclass of convex polynomials and, in particular, contains convex quadratic functions and convex separable polynomials. We also show that the relaxation problem can equivalently be reformulated as a semidefinite linear program.
Palabras clave / Keywords: robust optimization, SOS-convex polynomial, exact stable relaxation
Programado
Sesión GT11-3: Optimización Continua-3 (OPTIMIZACIÓN-3). Organizador: César Gutiérrez Vaquero
29 de mayo de 2018 12:20
Sala 6
