12.07.2015 Views

v2010.10.26 - Convex Optimization

v2010.10.26 - Convex Optimization

v2010.10.26 - Convex 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.

3.2. PRACTICAL NORM FUNCTIONS, ABSOLUTE VALUE 227A = {x∈ R 3 |Ax=b}R 3 B 1 = {x∈ R 3 | ‖x‖ 1 ≤ 1}Figure 70: 1-norm ball B 1 is convex hull of all cardinality-1 vectors of unitnorm (its vertices). Ball boundary contains all points equidistant from originin 1-norm. Cartesian axes drawn for reference. Plane A is overhead (drawntruncated). If 1-norm ball is expanded until it kisses A (intersects ballonly at boundary), then distance (in 1-norm) from origin to A is achieved.Euclidean ball would be spherical in this dimension. Only were A parallel totwo axes could there be a minimal cardinality least Euclidean norm solution.Yet 1-norm ball offers infinitely many, but not all, A-orientations resultingin a minimal cardinality solution. (1-norm ball is an octahedron in thisdimension while ∞-norm ball is a cube.)

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

Saved successfully!

Ooh no, something went wrong!