报告题目：Approximate KKT Reformulation and Necessary Optimality Conditions for Nonsmooth Bilevel Optimization Problems （非光滑双层优化问题的近似KKT转化与最优性必要条件）
报告人：万仲平 教授/博导 (武汉大学数学与统计学院)
摘要：In this paper, the bilevel programming problem is transformed into a single level programming problem by virtue of an approximate Karush-Kuhn-Tucker approach for the nonsmooth case. Close links with the bilevel programming problem for global or local optimal solutions are established. Particularly, the correspondences from solutions of the approximate KKT reformulation to ones of the bilevel programming problem are shown without the assumption of convexity. Moreover, a equivalent characterization of the M-type MPEC-MFCQ and two weaker constraint qualifications for nonsmooth optimization are provided. In addition, local optimal points of the nonsmooth approximate KKT reformulation converge to M-stationarity solutions of mathematical programs with complementarity constraints under the weaker constraint qualifications.