Why is it convex optimization

I see that Boyd ( of Stanford ) was visiting MIT and he has graciously made his book available on the web. I have gone through the lectures and prerequisites a couple times now and still can't get a complete enough grasp to extend the ideas to n-dimensional space and other problems. I intend to convert the mathematica matlab examples to python as a part of studying this again. I am combining this with the study of SAT axioms and proofs in general. Many of these ideas cross boundaries and concepts in signal processing, genetics or physics can show up in odd ways in the other sciences. This particular concept has been difficult for me to generalize ( actually even completely comprehend ) and apply in other ways so I am spending some time with it now.

I have an uneasy relationship with proofs and IIRC, somebody proved that a superconductor could not exist above 30°K, and it seems odd to even try to prove such a thing. Proving that something can never be done is a risky and unproductive endeavor. There is no product to apply. I suppose it is at least good to know in general that something is highly unlikely or impossible to avoid time on it, but it always interests me because it is like a mystery or something hidden. That doesn't mean that I would spend even a few seconds listening to a talk on perpetual energy machines.

if __name__ == "__main__":
 soup = BeautifulSoup(fo)
 links = SoupStrainer('a')
 k=[tag for tag in BeautifulSoup(fo, parseOnlyThese=links)]
 for child in soup.recursiveChildGenerator():
   name = getattr(child, "name", None)
  if name is not None:
     if name == 'a':
    aref= os.path.basename(child['href'])
    if  '.m' in aref:
     print aref
     wo = urllib.urlopen(weba+aref)
#     print fo

It doesn't seem to get the indentation right for tabs in blogger pre. So this is a start. I have made a script that parses the address of the matlab files and then downloads them to a place where they can be linked into zim for quick study and association to the equivalent Python substitute.

I think I am beginning to understand what convex optimization does and how it is applied. Odd thought, the outside of a neutron star is convex everywhere.

Don't do things that are already done. So:

aptitude search python | grep cvx
p   python-cvxopt   - Python package for convex optimization
v   python2.6-cvxopt-

And here is a link at UCLA to cvxopt.

Examples from the book Convex Optimization by Boyd and Vandenberghe

The above is a quote from the examples page for cvxopt at UCLA. Wow, it doesn't get any easier than this. Now all I have to do is understand what the heck it all means.

Okay now, on to stage 2. Since the examples are there, I can whip up a can of BeautifulSoup and urllib to correlate the examples from Boyd to the examples from (Joachim Dahl and Lieven Vandenberghe ) , then link them to my zim wiki with graphs and equations automatically generated. They even have a script to read and write matlab files. I wonder what it would have been like to grow up with such advantages that come from computer technology.

Unbelievable, they include graphics plotting and cross links to the book itself in the python code. On top of that they have an Open Office plug-in. They really did this up right, well done!

There is one CAVEAT in all this and it is "ez_install" and here is a discussion about using that for additions that aren't in debian. So you have been warned.

Just for reference at W3schools, I have had enough fun with this as autoplay. Alice Infinity discovered something about this and it will be a secret between me and her for the moment. There are so many ways around n-dimensional space that there are bound to be some short cuts through the twilight zone.


Automated Intelligence

Automated Intelligence
Auftrag der unendlichen LOL katzen