| Author |
Message
|
| rick.taiwan |
 Posted: Thu May 29, 2008 7:44 am Post subject: The Stochastic Trim-Loss Problem |
 |
|
Forum Freshman
Joined: 08 Mar 2008
Posts: 18
|
So here's a fundamental issue in resource management.
You've got a raw material -- wood, leather, whatever.
You need to cut it in order to manufacture stuff.
You don't want to overproduce, you don't want to lose profits.
However, demand is uncertain -- stochastic, in your model.
How do you stochastically program a combinatorially optimal solution?
Well, first you crack your Op Research book, then you look at the latest journals...
so I've been looking at EJOR - Euro Journal of Op Research.
They've got an article on this Trim-Loss problem, and I don't understand it entirely.
It's:
Beraldi, P., Bruni, M.E., Conforti, D., The Stochastic Trim-Loss Problem, European
Journal of Operational Research (2008), doi: 10.1016/j.ejor.2008.04.042
And it will take me some days to read it.
And I might get distracted.
But I might have to take a test on it at some point.
So I should read it.
Edit: Although, all things considered, I should probably re-read that other paper I have to hand in a report on first.
Edit:
If you don't intend to read the paper, let me tell you the cool part.
These guys programmed C++. But somehow they got LINDO as a callable library. I don't know how to do that, but I want to know how to do that.
And they said that there are lots of choices for callable optimization libraries.
That's what gets me going. That's motivating. IMHO. |
|
| Back to top |
|
 |
| rick.taiwan |
| Posted: Thu Jun 19, 2008 2:02 am Post subject: |
|
|
Forum Freshman
Joined: 08 Mar 2008
Posts: 18
|
| So I haven't coded this yet but from the lack of replies I'm guessing that everyone else finds the trim-loss problem underwhelming. |
|
| Back to top |
|
|
| jackson33 |
| Posted: Thu Jul 03, 2008 3:45 pm Post subject: |
|
|
Forum Professor
Joined: 10 Oct 2006
Posts: 1281
|
| Its not that cost/efficiency is not important to business, its VERY important, but that 'feasibility reporting' is very complicated. Think your concerns reflect sales, more so than cost to produce or efficiency. Its very rare for any new product to be produced by a new company, w/o some knowledge of a 'test marketing'. Its even more rare for a new product to be worried about efficiency over that marketing problem and usually priced well above cost to produce. If not a new product, another firm is the guide, but hard to compete, from scratch... |
|
| Back to top |
|
|
| rick.taiwan |
| Posted: Sat Jul 05, 2008 2:04 am Post subject: |
|
|
Forum Freshman
Joined: 08 Mar 2008
Posts: 18
|
| jackson33 wrote: |
| your concerns reflect sales, more so than cost to produce or efficiency. Its very rare for any new product to be produced by a new company, w/o some knowledge of a 'test marketing'. Its even more rare for a new product to be worried about efficiency over that marketing problem and usually priced well above cost to produce. |
True. Production is easier to model than sales, even stochastically -- and sales is the harder problem. I still need to figure out how they called LINDO from C++ and if it's an easy trick to do. |
|
| Back to top |
|
|
| maella |
| Posted: Mon Oct 06, 2008 12:31 am Post subject: |
|
|
Forum Freshman
Joined: 06 Oct 2008
Posts: 1
|
| Well, thinhs are really easier than you expect. Lindo API is a callable library exactly as CpLEX or whatever. Take a look on www.lindo.com -products-lindo api and enjoy! |
|
| Back to top |
|
|
|
|
