【转帖】need macro or software to optimize space in a bo

yang686526

need macro or software to optimize space in a box
hi all.....
i need a macro or software that can optimize orientation of many parts with different size and shape.
look at image below for more clear :
thanks for any advice
click for full image
a bit math intensive this one.
the basic principles are to get the box oriented at 0, 0, 0, 0, 0, 0 (x, y,z, pitch, yaw, roll) to make life much simpler. so you are working with a simple box size.
then loop all parts that are to go into the box and get their minimal bounding box. the easiest way is to do a brute force incremental analysis of rotating the part in all axis by increments of a degree or so and find out which orientation has the smallest box for an x, y and z of 0.
with all parts having their minimal bounding box size and required orientation it is then just a case of doing a basic math function to best fit squares into a square.
of course the minimal bounding box does not take into account complex shapes and will just box off all shapes, so if you were to do a load of shapes that looked like l-brackets for example then the code would be unsuitable.
if you want to get really complex and accurate your next step is to create nesting algorithms, but for more than one reason i am not going to go into detail about how to do that as i would be typing all day! lol
hi luke,
thanks for give me steps to solve this case...
unfortunately, i dont have programing background ?
do you know where i can by the program ?
or would you like to create those for me ? how about the price ?
thanks
i'd say me but at the moment i am extremely busy with other programs so the lead time would likely be a while but feel free to email me your requirements (complex nesting or simple) and any other info and i will liase with you on the matter (contact@angelsix.com)
this is essentially a unsolvable problem. because there's no way of verifying if a more optimum solution exist, except by iterating over all possible solutions. and that in essence is not possible. if the objects are box shaped then at least in theory its easier.
several vendors have software for this. theres actually wide variety of attacks against this. but as said its not easy to do a solve. theres atleast several very sophisticated thesis works done about this subject that i know of.
ps: by the way when and if somebody manages a solve each and every logistics/ sheetmetal cutting company etc... in the world will want to buy it form you.
there are many ways i could "solve" the issues by using the same 2d nesting algorithms such as dynamically adapting compact neighbourhood algo's (cna) or orthogonal meta-heuristic theories to work on a 2d plane (x,y), and a 2d plane (y,z) and the third (z,x), treat each as a 2d nest, then rank them together for best fit into the 3d space, and then use some heuristics to quickly chose the best ranking nests and eliminate 70%+ of lower success rate solutions. progress that to a 3 teir design and then for the successful teir do some actual trial and error cpu bashing to get actual results.
i didn't think there would be a mass market for such a thing but yes it is certainly possible, are there not already 3d nesting software packages out there?

edited: 10/29/2008 at 10:21 am by luke malpass
quick