Second-order optimality conditions with the envelope-like effect fornonsmooth vector optimization in infinite dimensionsSecond-order optimality conditions are of great interest, since they refine first-order conditions with second-orderinformation which is much helpful to recognize optimal solutions as well as to design numerical algorithms for computingthem. The nature of this second-order information is the following. Roughly speaking, first-order optimality conditions saythat at a minimizer the directional derivative of the map, composed from the objective and constraints, does not belongto the interior of the (composite) negative cone in the product of the image spaces. This directional derivative may lie onthe boundary of the mentioned cone. In this case, second-order conditions provide additional information. In general, thisinformation is that the second directional derivative of the Lagrangian is nonnegative. However, Kawasaki [1] discoveredthat, when one considers the closure of the mentioned negative cone, the second derivative of the Lagrangian may be strictlynegative if the encountered first-order directional derivative of the mentioned composite map lies on a particular part ofthe boundary of the mentioned negative cone. He called this phenomenon the envelope-like effect. A number of researchersstill do not pay attention to this effect, and some made mistakes related to this phenomenon. Many authors considerchỉ là hình nón phủ định được đề cập, không đóng cửa, và do đó không có phong bì như hiệu ứng xảy ra. Đã có đáng kể
đang được dịch, vui lòng đợi..