M. A. Melguizo Padial, F. García Castaño
This talk deals with a new Lagrange-type duality theory for constrained convex set-valued optimization problems defined between preordered normed spaces. We introduce a new set-valued dual program whose dual variables are pointed closed convex processes, and we derive from it both weak and strong duality results. The strong duality statement guarantees the existence of a dual solution under very weak assumptions (even if the optimal value in the primal program is unachieved). The approach is essentially geometric, and extends the methods used in the study of scalar programs to set-valued ones.
Keywords: Duality, Convex Programming, Process, Lagrange Multiplier, Set-valued
Scheduled
GT11 Continuous Optimization III
June 9, 2022 5:10 PM
A12