Decision Optimization

Hi all,

I want to reduce the size of decision variables, so it comes to me to use the script.

now I use : dvar boolean x[k in K][e in E];

but I know x[k][e] should be zero if e not in E[k], so how can I define x[k][e]?

I also try like below but failed.
dvar boolean x[k in K][e in E[k]]  ;

and 

execute {
for (var k in K)
  for( var e in E)
  { if (e not in E[k])
     x[k][e]=0; }  }

Thanks in advance!

------------------------------
Dario Pisinger
------------------------------

#DecisionOptimization

You can use a tuple (see the section about sparse modeling in the manual).
For example (untested code)
tuple KE {
  int k;
  int e;
}
{KE} ke = { <k,e> | k in K, e in E : e in E[k] };
dvar x[ke];

------------------------------
Daniel Junglas
------------------------------

Original Message:
Sent: Mon July 06, 2020 09:33 AM
From: Dario Pisinger
Subject: using script to define variables

Hi all,

I want to reduce the size of decision variables, so it comes to me to use the script.

now I use : dvar boolean x[k in K][e in E];

but I know x[k][e] should be zero if e not in E[k], so how can I define x[k][e]?

I also try like below but failed.
dvar boolean x[k in K][e in E[k]]  ;

and 

execute {
for (var k in K)
  for( var e in E)
  { if (e not in E[k])
     x[k][e]=0; }  }

Thanks in advance!

------------------------------
Dario Pisinger
------------------------------

#DecisionOptimization