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.

E.9. PROJECTION ON CONVEX SET 683<br />

The foregoing proof reveals another flavor of nonexpansivity; for each and<br />

every x,y ∈ R n<br />

‖Px − Py‖ 2 + ‖(I − P )x − (I − P )y‖ 2 ≤ ‖x − y‖ 2 (1916)<br />

Deutsch shows yet another: [92,5.5]<br />

E.9.4<br />

‖Px − Py‖ 2 ≤ 〈x − y , Px − Py〉 (1917)<br />

Easy projections<br />

Projecting any matrix H ∈ R n×n in the Euclidean/Frobenius sense<br />

orthogonally on the subspace of symmetric matrices S n in isomorphic<br />

R n2 amounts to taking the symmetric part of H ; (2.2.2.0.1) id est,<br />

(H+H T )/2 is the projection.<br />

To project any H ∈ R n×n orthogonally on the symmetric hollow<br />

subspace S n h in isomorphic Rn2 (2.2.3.0.1), we may take the symmetric<br />

part and then zero all entries along the main diagonal, or vice versa<br />

(because this is projection on the intersection of two subspaces); id est,<br />

(H + H T )/2 − δ 2 (H) .<br />

To project a matrix on the nonnegative orthant R m×n<br />

+ , simply clip all<br />

negative entries to 0. Likewise, projection on the nonpositive orthant<br />

R m×n<br />

−<br />

sees all positive entries clipped to 0. Projection on other orthants<br />

is equally simple with appropriate clipping.<br />

Projecting on hyperplane, halfspace, slab:E.5.0.0.8.<br />

Projection of y ∈ R n on Euclidean ball B = {x∈ R n | ‖x − a‖ ≤ c} :<br />

for y ≠ a , P B y = (y − a) + a .<br />

c<br />

‖y−a‖<br />

Clipping in excess of |1| each entry of a point x∈ R n is equivalent to<br />

unique minimum-distance projection of x on a hypercube centered at<br />

the origin. (conferE.10.3.2)<br />

Projection of x∈ R n on a (rectangular) hyperbox: [53,8.1.1]<br />

C = {y ∈ R n | l ≼ y ≼ u, l ≺ u} (1918)<br />

⎧<br />

⎨ l k , x k ≤ l k<br />

P(x) k=0...n = x k , l k ≤ x k ≤ u k (1919)<br />

⎩<br />

u k , x k ≥ u k

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

Saved successfully!

Ooh no, something went wrong!