Originally posted by: memo85

Hi,

I would like to solve the linear equations systems like the one given below by using CP. 

A unique solution of this equation systems is given below,

The solutions are continuous float values. But CP does not support the continuous variable. So, I did not model the equation system and solve it. How can I find a unique solution for any linear equation systems by using CP? Thank you in advance. 

Regards. 

Hi,

why not using CPLEX ?

dvar float l[1..3];

subject to
{
l[1]-0.1*l[2]-0.05*l[3]==5;
-0.75*l[1]+l[2]==0;
-0.25*l[1]-0.9*l[2]+l[3]==0;
}

gives your solution.

regards

Originally posted by: memo85

Hi, Alex. 

thank you very much for your reply. I do not think about using CPLEX. Your reply is really a helpful answer for me. Nevertheless, I wonder that whether it is possible by using CP, or not? 

Regards. 

Well then you could write

using CP;
 
 int scale=100000;
dvar int lscale[1..3] in -10*scale..10*scale;

dexpr float l[i in 1..3]=lscale[i]/scale;

subject to
{
abs(l[1]-0.1*l[2]-0.05*l[3]-5)<=0.0001;
abs(-0.75*l[1]+l[2])<=0.0001;
abs(-0.25*l[1]-0.9*l[2]+l[3])<=0.0001;
}

execute
{
writeln(l);
}

But I wonder why you do not want to use CPLEX.

Regards

Originally posted by: memo85

Thank you very much. Similar that equation system is part of my integrated model. And my model is CP based. So, firstly I think that I can model the equations by using CP. Also, I want to learn how to model continuous variable in CP. Using an integer scale is very advisable. 

Best regards.