Optimized cutting of master stock lengths using EXCEL with respect to minimising master stock 1

[MrZak]

Hey guys, need a bit of help or a point in the right direction here,

trying to create an excel spreadsheet to cut lengths out of a master stock with respect to minimising the master stock usage.

example:
i can get A lengths of Vmm

from this I need to cut:
B1 of Xmm
C1 of Ymm
D1 of Zmm etc.

I need to be able to calculate the best way in which to cut the master stock of Vmm to reduce the amount A
whilst also being able to input the width of the saw blade to take into account kerf (this will be kerf(mm) x n-1) n being the number of cuts made)

ANY help is greatly appreciated

 

[SkipVought]

Hi,

Welcome to Eng-Tips.

What have you tried so far? Please be specific and include the logic that you have tested and the results.

Are you cutting in one dimension, like lengths of extrusion, or two dimensions like pieces from a sheet? Your description seems to indicate the former.



Skip,
[sub]
Just traded in my OLD subtlety...
for a NUance![/sub]

 

[IRstuff]

Please do a search. This subject has been discussed at least a couple of time before.

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[MrZak]

IRstuff, I have searched for posts concerning this, of which you were the top poster! however, the excel spreadsheets I looked at in these posts seems to need the pattern solving to be done manually, or am I missing something here? Otherwise they have been really useful!

http://www.delphiforfun.org/Programs/Cutting Stock.htm#Download

which you posted seems to be pretty spot on the money, however being a mac user I can't run exe.

SkipVought, thankyou, I have found this a most helpful place indeed for the new work I am undertaking and look forward to learning more and more through this. Also, yes as indicated I am cutting from lengths in one dimension although once I have completed this I will also be looking at 2d sheetwork but I haven't even begun the research on this! On paper I can solve the problem using bin packing LP and IP for upper and lower bounds, however upon researching here there seems to be a bit more and transcribing it into a usable programme is a bit beyond me.

 

[SkipVought]

MrZak,

No doubt you've googled this issue. Nesting optimization is not a trivial matter. I have worked with a colleague who actually has a patent pending doe a nesting process: from his LinkedIn; "Developed the “Nesting” concept for raw material and detail fabrication process management. The ‘nesting’ process has a patent pending."

You your looking for a "freebie" is gonna cost you in one way or another.

Skip,
[sub]
Just traded in my OLD subtlety...
for a NUance![/sub]

 

[IRstuff]

Depending on the Mac, some of them can run DOS or Windows programs using the Wine emulator, or running Windows in a virtual machine on Mac, or dual booting Windows and Max

TTFN
I can do absolutely anything. I'm an expert!
faq731-376 forum1529

 

[MrZak]

Cheers IR, managed to boot the programme using wine emulator before I read this, the programme is useful but for now actually more than I want!
Skipvought, not looking for a freebie just interested in what maths people are employing for this, I think you think I want more than I do! The programme I'm looking for exists as free download in a lot of places. I just wanted to try and engineer one for myself, more project than business but it'll also come in handy at work!

 

[zelgar]

Part of the problem is that the most efficient cutting strategy that creates the least amount of scrap from the stock may result insufficient amounts of pieces. (e.g., let's say the most efficient cutting strategy gets you 10 pieces of size X, 2 pieces of size Y and 5 pieces of size Z; but your widget needs 2 size X, 5 size Y and 1 size Z.) Your cutting strategy needs to be based upon your needs, not just the most number of pieces you can get from a master stock sized V.

 

[SkipVought]

How about a well defined logic specification. You intimated that you were using some logic, but never stated it explicitly.

Skip,
[sub]
Just traded in my OLD subtlety...
for a NUance![/sub]

 

[djs]

Not a MAC program, but a Windows program check out MaxCut. Works great!

 

[SkipVought]

IN all fairness, the OP has stated...

i can get A lengths of Vmm

from this I need to cut:
B1 of Xmm
C1 of Ymm
D1 of Zmm etc.

There are his needs. We have Supply (A lengths of Vmm) and Demand (B1 of Xmm, C1 of Ymm, D1 of Zmm etc.), and the desire to minimize the Supply for a given Demand.

Now he needs to lay out his nesting strategy/logic in order that it can be appropriately coded in Excel, which is where we can come in and help.

Skip,
[sub]
Just traded in my OLD subtlety...
for a NUance![/sub]

 

[zelgar]

Here's a little spreadsheet with a macro that will calculate where you need to make cuts. It's currently set up for having 3 sizes (e.g., x, y and z) and assumes the largest piece is x and down to z being the smallest. The results may not be the most efficient cutting method, but it will calculate where the cuts need to begin and verify that you have enough stock in order to do the job.

<edited> Original spreadsheet used the Long data type for non-integers. Also, when remaining stock was less than the cut width would result in a negative value.

 

[SkipVought]

Still waiting on MrZak's logic.

Skip,
[sub]
Just traded in my OLD subtlety...
for a NUance![/sub]

 

[zelgar]

Even though we haven't heard back from the OP, I modified my original spreadsheet to do up to 10 different lengths. I won't state that it would result in the most efficient cuts, but it does try to limit the amount of scrap by maximizing the number of cuts by making the most cuts of the bigger lengths down to the smallest. Additionally, the cut sizes don't need to be put in order by size (the macro will resort the lengths first).

Note: to try to optimize the smallest amount of scrap, you'd need to figure out all of the different combinations of the cuts on all of the different boards. Once you found all of the different combinations, then you'd need to calculate the amount of scrap was left over.

 

[TehMightyEngineer]

zelgar, that's a nice idea; it's not perfect optimization but is automated, easy to use, and will definitely be more efficient than just cutting X pieces of piece #1, Y pieces of piece #2, etc. This gave me the inspiration to finally get around to starting work on a rebar cutting/bending optimization program for our precast pieces.

Professional and Structural Engineer (ME, NH, MA)
American Concrete Industries

http://www.americanconcrete.com

 

[zelgar]

The biggest problem with trying to find the optimal solution is that the number of combinations get big really fast depending upon the both the number of different types of pieces and how many pieces of each type could be cut from the stock. I still have been messing around with the macros to see if I could get a more efficient solution, but the number of combinations is making things difficult.

 

[TehMightyEngineer]

I'm honestly wondering if it might be worth it to set it up with a random number generator so that it tries random combinations and let it brute force the cut combinations. You could then have it compare each iteration to the last and if the last X number of attempts resulted in no improvement then it stops. X could be set to be some number of combinations that ran in a reasonable amount of time and thus it would balance out finding an ideal solution with simplicity of programing and computing time required. You could also give it a few logic checks so that it provides some order to the random combinations and thus avoids obviously poor choices.

This would likely be a reasonable approximation of the ideal solution.

Professional and Structural Engineer (ME, NH, MA)
American Concrete Industries

http://www.americanconcrete.com