Is it possible to solve a 2d bin packing problem or nesting problem with Cplex? The solution should present the $x$ and $y$ coordinates of each item.
$\begingroup$
$\endgroup$
2
-
1$\begingroup$ please accept the answer that solved your problem rather than changing the question to "solved". $\endgroup$– EhsanK ♦Feb 3, 2021 at 16:49
-
$\begingroup$ @Mathew, beside of the interested answer of Alex, you might want to take a look into the ABS models to solve floor layout problems. :) $\endgroup$– A.OmidiFeb 4, 2021 at 8:14
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
you could start with
//using CP;
int n=25; // nb of rectangles
int sizeBigSquare=10000;
int l[i in 1..n]=rand(100);
int w[i in 1..n]=rand(100);
dvar int x[i in 1..n]in 0..sizeBigSquare;
dvar int y[i in 1..n] in 0..sizeBigSquare;
dvar boolean rota[i in 1..n]; // do we rotate?
dexpr int x2[i in 1..n]=x[i]+(rota[i]==1)*w[i]+(rota[i]==0)*l[i];
dexpr int y2[i in 1..n]=y[i]+(rota[i]==1)*l[i]+(rota[i]==0)*w[i];
dexpr int Mx=max(i in 1..n) x2[i];
dexpr int My=max(i in 1..n) y2[i];
minimize My;
subject to
{
Mx<=1000;
forall(ordered i,j in 1..n)
(x2[i]<=x[j]) || (x2[j]<=x[i]) || (y2[i]<=y[j]) || (y2[j]<=y[i]);
}
float rate=sum(i in 1..n) w[i]*l[i]/My/Mx;
execute
{
writeln("Rate=",rate);
}
that is relying on MIP within CPLEX
And if you comment // in //using CP; then this model will use constraint programming
-
$\begingroup$ Thanks for your interesting solution. Would you please, add some details in your answer to clarify decision variables and constraints notation? $\endgroup$– A.OmidiFeb 4, 2021 at 8:18