A simple characterization of solutions sets of convex programs (Q1108196): Difference between revisions

The author shows that for any convex program \(\min_{x\in X}f(x)\), where X is a convex set in R n and f(x) is a convex function on R n, the subdifferential \(\partial f(x)\) is constant on the relative interior of the set \(\bar X=\arg \min_{x\in X}f(x)\) (solution set of the problem) and equals the intersection of the subdifferentials of the function f(x) at all points of \(\bar X.\) In addition, \(\bar X\) lies in the intersection with the feasible set X of an affine subspace orthogonal to some subgradient of f(x) at a relative interior point of \(\bar X.\) As a consequence a simple polyhedral characterization is given for the solution set of a convex quadratic program and that of a monotone linear complementarity problem.