10.03.2015 Views

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

2.13. DUAL CONE & GENERALIZED INEQUALITY 157<br />

A T y ≺ 0<br />

or in the alternative<br />

Ax = 0, x ≽ 0, ‖x‖ 1 = 1<br />

<br />

(318)<br />

2.13.3 Optimality condition<br />

The general first-order necessary and sufficient condition for optimality<br />

of solution x ⋆ to a minimization problem ((275p) for example) with<br />

real differentiable convex objective function f(x) : R n →R is [265,3]<br />

(confer2.13.10.1) (Figure 59)<br />

∇f(x ⋆ ) T (x − x ⋆ ) ≥ 0 ∀x ∈ C , x ⋆ ∈ C (319)<br />

where C is a convex feasible set (the set of all variable values satisfying the<br />

problem constraints), and where ∇f(x ⋆ ) is the gradient of f (3.1.8) with<br />

respect to x evaluated at x ⋆ . 2.58<br />

In the unconstrained case, optimality condition (319) reduces to the<br />

classical rule<br />

∇f(x ⋆ ) = 0, x ⋆ ∈ domf (320)<br />

which can be inferred from the following application:<br />

2.13.3.0.1 Example. Equality constrained problem.<br />

Given a real differentiable convex function f(x) : R n →R defined on<br />

domain R n , a fat full-rank matrix C ∈ R p×n , and vector d∈ R p , the convex<br />

optimization problem<br />

minimize f(x)<br />

x<br />

(321)<br />

subject to Cx = d<br />

is characterized by the well-known necessary and sufficient optimality<br />

condition [53,4.2.3]<br />

∇f(x ⋆ ) + C T ν = 0 (322)<br />

where ν ∈ R p is the eminent Lagrange multiplier. [264] Feasible solution x ⋆<br />

is optimal, in other words, if and only if ∇f(x ⋆ ) belongs to R(C T ).<br />

2.58 Direct solution to (319) instead is known as a variation inequality from the calculus<br />

of variations. [215, p.178] [110, p.37]

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!