Originally posted by: bl4ckic3

Hi all
I would like to know how to use OPL to model such equation:

http://bl4ckic3.persiangig.com/opl.png

The problem I faced in the pdf manual for OPL is they never show you mathematical model so I don't know how they designed the code.
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

Hello.

It's very simple. Just use "forall", "sum" and "max". Lets take for example (1). Assuming that Q and Qmax are integer variables and OB is an integer constant, the problem can be modeled as follows:

int d = 3; 

int n = 3; dvar 

int Qmax[1..d][1..d]; dvar 

int Q[1..d][1..d][1..n]; 

int OB = 10;   constraints 
{ forall(b in 1..d) OB == sum(f in 1..d, j in 1..n) (Qmax[b][f] - Q[b][f][j]); 
}

Best, Petr
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: bl4ckic3

Dear Peter, thank you for your awesome response, I have better understanding now.
I am completely beginner in this subject and I have very limited time
I started to write constraint.
the first one :

forall(b in 1..d)
Qmax[b][f]== max(Q[b][f][j]);
write Qmax[b][f];

I don't know its right to ask such stupid question here, but I wanted to know is it right way for defining first constraint?
#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

Hello.

This equation is not clear in your picture: it doesn't say on what index operator max iterates. Is it j? Assuming that it is j, the right syntax in OPL is:

forall(b in 1..d) forall(f in 1..d) Qmax[b][f] == max(j in 1..m) Q[b][f][j];

Petr
#DecisionOptimization
#OPLusingCPOptimizer