Global infimum of strictly convex quadratic functions with bounded perturbations

Minimizing Convex Functions with Bounded Perturbations

“…In [19] we showed that (2.5) h 1 (γ) = 1 4 λ min γ 2 for γ ∈ [0, +∞[, where λ min is the smallest eigenvalue of A.…”

“…But we can get better upper bounds for specific problem classes. For instance, for quadratic function f given in (2.4), it is shown in [19] that this upper bound is γ * /2.…”