Mathematics of Operations Research · 2005 · 32 citations · 8 references
Natural SciencesMathematicsDerivativesFormal MethodsOptimizationBoundary ConditionsBoundary-value ProblemsConvex AnalysisMetric GeometryConstraint QualificationsConvex FunctionsProper FunctionRegularity TheoryStrong ConvexityStrong Duality
In this paper, we characterize the general difference between strong basic constraint qualification (BCQ) and BCQ. For this purpose, we introduce a new measurement, the extent of a subdifferential, and show that for an inequality defined by a proper convex function f, the strong BCQ at a boundary point x of the solution set is equivalent to the extended BCQ plus the positivity of the extent of subdifferential at x. Applying the above characterization to the case when f is the maximum of finitely many differentiable convex functions, we show that the metric regularity at a boundary point x is equivalent to BCQ at every point in a “boundary-neighborhood” of x. In addition, we provide an answer to the open question proposed by Zheng and Ng [ 11 ]. We construct an example to show that BCQ at a boundary point x does not ensure the metric regularity at x.
8
Error bounds in mathematical programming
Jong‐Shi Pang · Mathematical Programming · 1997 · 451 citations