On Numerically Verifiable Exactness of Multiplier Relaxations
Abstract
Robust semi-definite programming problems with rationaldependence on uncertainties are known to have a wide rangeof applications, in particular in robust control.It is well-established how to systematically constructrelaxations on the basis of the full block S-procedure. Ingeneral such relaxations are expected to be conservative,but for concrete problem instances they are often observedto be tight. The main purpose of this paper is to investigatedin how far recently suggested tests for the exactnessof such relaxations are indeed numerically verifiable.