|
|
 | | From: | Tom | | Subject: | Difference between Dynamic Programming and Combinatorics | | Date: | 4 Jan 2005 13:36:43 -0800 |
|
|
 | Could someone describe to me, in plain words, the difference between
these two approaches while solving a non-convex optimization problems.
I am inclined to think that they would behave in a similar fashion.
The optimization problem involoves both continuous and discrete
variables.
Thanks.
Tom
|
|
| | From: | Roger Bagula | | Subject: | Re: Difference between Dynamic Programming and Combinatorics | | Date: | Fri, 14 Jan 2005 15:42:12 GMT |
|
|
| Dear Tom,
I wish you good luck in finding someone who can answer that question.
I've done work in fractal combinatoial problems .
They are mostly "integer" even when they give the same type solution as
a "map": Pascal's triangle modulo 2 and Sierpinski gasket
are examples.
Most chaotic dynamics isn't in "map" types, but in differential
equations ( with exceptions like the fuzzy, Henon, and Lozi maps).
By optimization problems you lost most of us. It seems to be a specific
kind of engineering problem ( mostly chemical plant engineering).
The old fashioned way to model such systems was to use linear equations
which are closer to hydrodynamics than to chaotic dynamics.
It is general systems theory where this stuff does overlap ( biological
systems/ closed, flames / open). My impression of the opimization stuff
on the web is they renamed it without referencing the older material so
students are confused by the naming.
Esssentially you are trying to get a system model that has equations
that behave as your system does. The optimization comes afterwards
when you are trying to get the smallest set of equations that behave
like the system.
My suggestion is that you ask an engineering professor:
http://titan.princeton.edu/home.html
Tom wrote:
> Could someone describe to me, in plain words, the difference between
> these two approaches while solving a non-convex optimization problems.
> I am inclined to think that they would behave in a similar fashion.
> The optimization problem involoves both continuous and discrete
> variables.
>
> Thanks.
>
> Tom
|
|
|
