User talk:Wotao.yin

From Wikimization

(Difference between revisions)
Jump to: navigation, search
Line 23: Line 23:
Vector <math>c\,</math> may arbitrarily be set to <math>\mathbf{0}</math> or <math>\mathbf{1}</math>, for your purposes, but leave a hook for it in case you require another value.
Vector <math>c\,</math> may arbitrarily be set to <math>\mathbf{0}</math> or <math>\mathbf{1}</math>, for your purposes, but leave a hook for it in case you require another value.
-
A good presolver can eliminate about 50,000 columns of <math>\,A</math> because one of the constraints '''('''fifth row from the bottom of <math>\,A\,</math>''')''' has only nonnegative entries. &nbsp; This means that about 50,000 entries in permutation submatrix <math>X\,</math> can be set to zero before numerical solution begins. &nbsp; The [http://www.convexoptimization.com/TOOLS/Wotao.Yin/WotaoX.mat Matlab binary] possesses all 262,144 columns of <math>A\,</math>; &nbsp; none of its columns have been discarded in a presolve.
+
A good presolver can eliminate about 50,000 columns of <math>\,A</math> because one of the constraints '''('''fifth row from the bottom of <math>\,A\,</math>''')''' has only nonnegative entries. &nbsp; This means that about 50,000 entries in permutation submatrix <math>X\,</math> can be set to zero before numerical solution begins. &nbsp; The [http://www.convexoptimization.com/TOOLS/Wotao.Yin/WotaoX.mat Matlab binary] possesses all 262,144 columns of <math>A\,</math>; &nbsp; none of its columns have yet been discarded by a presolve.
--[[User:Dattorro|Dattorro]] 03:31, 5 November 2010 (PDT)
--[[User:Dattorro|Dattorro]] 03:31, 5 November 2010 (PDT)

Revision as of 19:19, 5 November 2010

I regard the following as a very difficult problem, having spent considerable time with it.

LaTeX: \begin{array}{cl}\mbox{minimize}_X&c^{\rm T}\mbox{vec}\,X\\
\mbox{subject to}&A\,\mbox{vec}\,X=b\\
&X^{\rm T\!}X=I\\
&X\geq_{}\mathbf{0}\end{array}

Rectangular submatrix LaTeX: \,X\!\in\mathbb{R}^{1024\times256}\, comes from a permutation matrix LaTeX: \,\Xi\!\in\!\mathbb{R}^{1024\times1024}\, having three out of every four consecutive columns discarded.   This discard occurs because of structural redundancy in LaTeX: \Xi\,.

Notation LaTeX: \mbox{vec}\,X denotes vectorization; it means, the columns of LaTeX: \,X are stacked with column 1 on top and column 256 on the bottom.

Matrix LaTeX: A\!\in\!\mathbb{R}^{10565\times262144} is sparse having only 979,444 nonzeros.   It contains only integers from the set LaTeX: \{{-1},0,1,2\}\,.  The 2 appears only in the fifth row from the bottom of LaTeX: A\,.

Vector LaTeX: b\, is quite sparse having only a single nonzero entry: LaTeX: 1\,.

A Matlab binary contains matrices LaTeX: \,A and LaTeX: \,b.   Vector LaTeX: c\, is left unspecified because I want to vary it later as part of a Convex Iteration.   Vector LaTeX: c\, may arbitrarily be set to LaTeX: \mathbf{0} or LaTeX: \mathbf{1}, for your purposes, but leave a hook for it in case you require another value.

A good presolver can eliminate about 50,000 columns of LaTeX: \,A because one of the constraints (fifth row from the bottom of LaTeX: \,A\,) has only nonnegative entries.   This means that about 50,000 entries in permutation submatrix LaTeX: X\, can be set to zero before numerical solution begins.   The Matlab binary possesses all 262,144 columns of LaTeX: A\,;   none of its columns have yet been discarded by a presolve.

--Dattorro 03:31, 5 November 2010 (PDT)

Personal tools