Robust linear semi-infinite programming duality under uncertainty (Q353146)

In this paper duality is studied for linear semi-infinite programming (LSIP) problems which involve uncertainty in the constraint functions, in the objective function, or in both, where this is done within the framework of robust optimization. In particular, robust duality is established for such an uncertain LSIP problem via the proof of strong duality between the robust counterpart of an uncertain LSIP problem and the optimistic counterpart of its uncertain Lagrangian dual. It is furthermore shown that robust duality holds if the robust moment cone is closed and convex and that, conversely, the latter condition implies the presence of robust duality in case of uncertainty in the constraints. In addition, characterizations of robust versions of the Farkas lemma for infinite linear inequality systems under data uncertainty are provided.