Strong Abadie CQ, ACQ, calmness and linear regularity

• 成果类型: Article
• 署名作者: Wei, Zhou; Yao, Jen-Chih; Zheng, Xi Yin
• 署名单位: Yunnan University; Kaohsiung Medical University; National Sun Yat Sen University
• 刊物名称: MATHEMATICAL PROGRAMMING
• ISSN/ISSBN: 0025-5610
• DOI: 10.1007/s10107-013-0641-4
• 发表日期: 2014
• 页码: 97-131
• 关键词: weak sharp minima; error-bounds; constraint qualifications; generalized equations; metric subregularity; strong chip; convex; intersection; MULTIFUNCTIONS; Duality

摘要

The Abadie CQ (ACQ) for convex inequality systems is a fundamental notion in optimization and approximation theory. In terms of the contingent cone and tangent derivative, we extend the Abadie CQ to more general convex multifunction cases and introduce the strong ACQ for both multifunctions and inequality systems. Some seemly unrelated notions are unified by the new ACQ and strong ACQ. Relationships among ACQ, strong ACQ, basic constraint qualification (BCQ) and strong BCQ are discussed. Using the strong ACQ, we study calmness of a closed and convex multifunction between two Banach spaces and, different from many existing dual conditions for calmness, establish several primal characterizations of calmness. As applications, some primal characterizations for error bounds and linear regularity are established; in particular, some existing results are improved.