31.12.2013 Views

Lecture 2 Piecewise-linear optimization

Lecture 2 Piecewise-linear optimization

Lecture 2 Piecewise-linear optimization

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Proof of nullspace condition<br />

notation<br />

• x has support I ⊆ {1,2,...,n} if x i = 0 for i ∉ I<br />

• |I| is number of elements in I<br />

• P I is projection matrix on n-vectors with support I: P I is diagonal with<br />

(P I ) jj =<br />

{<br />

1 j ∈ I<br />

0 otherwise<br />

• A satisfies the nullspace condition if<br />

‖P I z‖ 1 < 1 2 ‖z‖ 1<br />

for all nonzero z in nullspace(A) and for all support sets I with |I| ≤ k<br />

<strong>Piecewise</strong>-<strong>linear</strong> <strong>optimization</strong> 2–17

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

Saved successfully!

Ooh no, something went wrong!