[AMPL 2887] Solution accuracy

[Saber Jalilzadeh-Galaeh]

Some parts of this message have been removed. Learn more about Nabble's security policy.

Dear All Users,

I am using AMPL/CPLEX and working on a LP problem.
One of my solutions is a table as:

 

x[1,1,3,2,*,*]
:   1        2        3        4             5         6        7        8   :=
1   0   0             0   0              0             0   0             0
2   0   0             0   0             -7.14503e-13   0   6.61394e-13   0
3   0   0             0   0              0             0   0             0
4   0   0             0   8.64909e-13    1             0   0             0
5   0   0             0   0              0             0   0             0
6   0   0             0   0             -1.23051e-12   0   0             0
7   0   0             0   0              0             0   3.18295e-13   0
8   0   1.23051e-12   0   0              0             1   0             0

 

As you see, there are very small numbers which
I want to be equivalent with zero.
I mean that if the solution is less that 1e-3 then CPLEX replace it with zero.
It is equivalent with imposing the solution accuracy with 1e-3.

Would you please help me with which command I could do that?

Thanks in advance
Saber

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---

[Ali Baharev]

> I mean that if the solution is less that 1e-3 then CPLEX replace it with
> zero.

You can only ask AMPL to show them as if they were zero. Just add the
corresponding condition to the display command.

Good luck,

Ali

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---

[Robert Fourer-2]

Some parts of this message have been removed. Learn more about Nabble's security policy.

You can use option solution_round 3 to round all solution values received from CPLEX at the third place after the decimal point, thus forcing all those tiny numbers to zero.  Or you can use option display_eps 1e-3 to display all numbers less than .001 in magnitude as zero; the difference in this case is that the actual values aren't changed, only the way they are displayed.

 

Bob Fourer

4er@...

 

 

---

From: ampl@... [mailto:ampl@...]

On Behalf Of Saber Jalilzadeh-Galaeh [sbr_612000@...]
Sent: Monday, October 19, 2009 7:05 AM
To: AMPL Group
Subject: [AMPL 2887] Solution accuracy

 

Dear All Users,

I am using AMPL/CPLEX and working on a LP problem.
One of my solutions is a table as:

 

x[1,1,3,2,*,*]
:   1        2        3        4             5         6        7        8   :=
1   0   0             0   0              0             0   0             0
2   0   0             0   0             -7.14503e-13   0   6.61394e-13   0
3   0   0             0   0              0             0   0             0
4   0   0             0   8.64909e-13    1             0   0             0
5   0   0             0   0              0             0   0             0
6   0   0             0   0             -1.23051e-12   0   0             0
7   0   0             0   0              0             0   3.18295e-13   0
8   0   1.23051e-12   0   0              0             1   0             0

 

As you see, there are very small numbers which
I want to be equivalent with zero.
I mean that if the solution is less that 1e-3 then CPLEX replace it with zero.
It is equivalent with imposing the solution accuracy with 1e-3.

Would you please help me with which command I could do that?

Thanks in advance
Saber

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---

[Saber Jalilzadeh-Galaeh]

Some parts of this message have been removed. Learn more about Nabble's security policy.

Dear Robert, Paul and Ali,

 

Thanks a lot for your helps.

 

 

Saber

---

From: Robert Fourer <4er@...>
To: ampl@...; Saber Jalilzadeh-Galaeh <sbr_612000@...>
Sent: Mon, October 19, 2009 8:55:10 PM
Subject: [AMPL 2892] Re: Solution accuracy

You can use option solution_round 3 to round all solution values received from CPLEX at the third place after the decimal point, thus forcing all those tiny numbers to zero.  Or you can use option display_eps 1e-3 to display all numbers less than .001 in magnitude as zero; the difference in this case is that the actual values aren't changed, only the way they are displayed.

 

Bob Fourer

4er@...

 

 

---

From: ampl@... [mailto: ampl@... ]

On Behalf Of Saber Jalilzadeh-Galaeh [sbr_612000@...]
Sent: Monday, October 19, 2009 7:05 AM
To: AMPL Group
Subject: [AMPL 2887] Solution accuracy

 

Dear All Users,

I am using AMPL/CPLEX and working on a LP problem.
One of my solutions is a table as:

 

x[1,1,3,2,*,*]
:   1        2        3        4             5         6        7        8   :=
1   0   0             0   0              0             0   0             0
2   0   0             0   0             -7.14503e-13   0   6.61394e-13   0
3   0   0             0   0              0             0   0             0
4   0   0             0   8.64909e-13    1             0   0             0
5   0   0             0   0              0             0   0             0
6   0   0             0   0             -1.23051e-12   0   0             0
7   0   0             0   0              0             0   3.18295e-13   0
8   0   1.23051e-12   0   0              0             1   0             0

 

As you see, there are very small numbers which
I want to be equivalent with zero.
I mean that if the solution is less that 1e-3 then CPLEX replace it with zero.
It is equivalent with imposing the solution accuracy with 1e-3.

Would you please help me with which command I could do that?

Thanks in advance
Saber

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---