Lecture I (Days 2,3,4):
Convex Analysis
(
Back to Main Course Page;
back to Lecture 0;
onto Lecture 2;
)
-
affine set, convex set, convex combinations and hull, convex cone,
hyperplanes and halfspaces, Euclidean balls and ellipsoids, norm
balls and norm cones, polyhedra, positive semidefinite cone, minimum
and minimal elements, separating and supporting hyperplane theorems
(and proofs), dual cones
Assignments:
-
Derive necessary (and sufficient) conditions for positive
definiteness of X in Sn based on the angle between X and I.
-
Prove that K=K++ iff K is a c.c.c.
-
first and second order conditions, epigraph, Jensen's inequality,
compositions, pointwise supremum, minimization, conjugate function
Assignments:
-
Prove that a real valued convex function on Rn is continuous.
-
Prove that a real valued convex function on Rn is
locally Lipshitz continuous.
-
Prove that the function G(X)=XXT-I, from
n by n matrices to n by n symmetric matrices, is K-convex with K the
cone of semidefinite matrices.
-
Prove that the function g(X)=log det (X) is strictly concave on the set
of n by n positive definite matrices X. Find the gradient and Hessian of
g(X).
-
standard form, optimal points, implicit constraints, feasibility,
(geometric) optimality conditions, quasiconvex optimization,
fractional programs, quadratic programs, quadratically constrained
quadratic programs, second-order cone programs, robust linear
programming, geometric programming, semidefinite programming,
eigenvalue minimization, multicriteria minimization, portfolio
optimization
-
Lagrangian, Lagrange dual function, Lagrange dual problem, weak and
strong duality!!, constraint qualifications, KKT optimality
conditions, dual SDP
Supplementary Information
-
-
Suggested Texts/References:
Back to Main Course Page
Start of
's
Home Page
(more
ambigrams)
Users' Home Pages on
orion.math,