Semidifferential of Minimax of Lagrangians with Respect to Parameters, Controls, or Geometric Variables
Michel C. Delfour · SIAM Journal on Control and Optimization · 2025
Abstract. The object of this paper is to further investigate the use of the minimax of a parametrized Lagrangian to get the semidifferential, that is, the one-sided directional derivative, of objective functions constrained by variational equations or inequalities, which depend on parameters, controls, or the shape/topology of the underlying geometric domain. It extends earlier results to Lagrangians with or without saddle point arising from state constrained objective functions.