[AMPL 2576] Assignment and constraints

[x-22]

Hi all,

I have the following. I'm keeping the syntax simple here for ease of discussion.

var L{0..4}, M{0..4};
var final >= 0;

c1{j in 0..4}: L[j] <= M[j];

c2{j in 1..4}: L[j] = L[0];

c{j in 0..4}: final = sum{j} (M[j] - L[j]);

minimize{j in 0..4}: final;

#need this explicitly?
#c3{j in 0..4}: L[0] <= M[j];

This crashes FilMINT solver. Unfortunately, I don't get any indication of what or where the problem is when the solver crashes.. So, I tried it with MINOS solver; it ignores integrality (not related to the variables shown here) but solves the problem. However, it picks L[0] to have a value that is not <= each M[j] which makes final negative. But this is not what I expected.

So, my question is this - when L[1..4] is assigned L[0], will L[0] value be selected by the solver to be <= each of the M[j] per constraint c1?

[Or, does the solver pick L[0] <= M[0], assign L[0] to the rest of the L[j] and then crash if c1 or final_LB is violated? In that case, I'd have to include c3 constraint but even that hasn't helped me with FilMINT so far. However, me forcing L[0] to have a valid value (through a constraint) helps..]

Sincere thanks in advance for any and all information!


     

--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---

[stefano.coniglio]

> minimize{j in 0..4}: final;

With this declaration you have 5 objective functions. Which one are
you minimizing?

--~--~---------~--~----~------------~-------~--~----~
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]

When you write "c1{j in 0..4}: L[j] <= M[j]" AMPL sets j successively to the
values 0, 1, 2, 3, 4 and generates the corresponding constraints L[0] <=
M[0], L[1] <= M[1], L[2] <= M[2], L[3] <= M[3], L[4] <= M[4].  

Evidently you're looking for something else, so you need to write your
constraint differently.  The constraint "c3{j in 0..4}: L[0] <= M[j]" seems
more nearly like what you are describing, since it generates L[0] <= M[0],
L[0] <= M[1], L[0] <= M[2], ...

To be sure you could get a listing of all the constraints generated by
giving the AMPL commands "expand c1;", "expand c3;" and the like.  The
solver does not have any effect on what constraints AMPL generates; only
after AMPL generates all the constraints (and the objective) does the solver
try to find a solution for them.  If FilMINT is crashing on some problem
then you best bet is to bring that to the attention of FilMINT's developers.

Bob Fourer
4er@...

> -----Original Message-----
> From: ampl@... [mailto:ampl@...] On Behalf Of x
> Sent: Thursday, June 04, 2009 1:16 PM
> To: ampl@...
> Subject: [AMPL 2576] Assignment and constraints
>
>
>
> Hi all,
>
> I have the following. I'm keeping the syntax simple here for ease of
> discussion.
>
> var L{0..4}, M{0..4};
> var final >= 0;
>
> c1{j in 0..4}: L[j] <= M[j];
>
> c2{j in 1..4}: L[j] = L[0];
>
> c{j in 0..4}: final = sum{j} (M[j] - L[j]);
>
> minimize{j in 0..4}: final;
>
> #need this explicitly?
> #c3{j in 0..4}: L[0] <= M[j];
>
> This crashes FilMINT solver. Unfortunately, I don't get any indication of
> what or where the problem is when the solver crashes.. So, I tried it with
> MINOS solver; it ignores integrality (not related to the variables shown
> here) but solves the problem. However, it picks L[0] to have a value that
> is not <= each M[j] which makes final negative. But this is not what I
> expected.
>
> So, my question is this - when L[1..4] is assigned L[0], will L[0] value
> be selected by the solver to be <= each of the M[j] per constraint c1?
>
> [Or, does the solver pick L[0] <= M[0], assign L[0] to the rest of the
> L[j] and then crash if c1 or final_LB is violated? In that case, I'd have
> to include c3 constraint but even that hasn't helped me with FilMINT so
> far. However, me forcing L[0] to have a valid value (through a constraint)
> helps..]
>
> Sincere thanks in advance for any and all information!
>

--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---