[ExI] mersenne primes again

Mike Dougherty msd001 at gmail.com
Fri Apr 2 00:06:20 UTC 2010


On Thu, Apr 1, 2010 at 5:35 PM, spike <spike66 at att.net> wrote:
> the go-to guy on math; any math question I had the reputation as the one to
> see, and please do, I love all kinds of math.  But closed-form integration

'made me think to ask you...

> Recently Keith and I were fooling with a spreadsheet I wrote about 15 yrs
> ago to estimate rocket booster performance taking into account aerodynamic
> drag and rocket motor thrust.  That is a really good example of a real-world
> problem that cannot be done effectively in closed form, because there are
> too many variables changing too many ways.

... and this is a nice segue ...

> Better than spending a year on calculus would be a parallel path which
> students could choose, which takes advantage of a tool we didn't have in my
> misspent youth: the spreadsheet.

Have you read anything about Genetic Algorithms as a solution finder?

I'm working on an undergrad seminar paper that discusses a GA
implemented in SQL for the purpose of discovering a relationship
between keywords in a person's business title and their likelihood of
clicking on a link in an email.

The beauty of the genetic algorithm is that it finds 'pretty good'
solutions even when the ideal solution is not known and when there are
too many interdependent variables to make other solution finders
perform in reasonable time (as in your rocket design problem)  The
basic design is fairly easy to implement.  The rest of exploration and
exploitation of the solution space is done by throwing CPU/resource at
it.  That wasn't too feasible 30 years ago, but it's pretty affordable
today.  I was mostly curious if you've even heard of genetic
algorithms...




More information about the extropy-chat mailing list