How to deal with impossible deduced bounds for a variable?

[vinhk...@gmail.com]

Hi, 

I am running a code to visit each waypoint from a list of options. 

%sample waypoints

V = [520.0000 -200.0000;

 -961.9101  314.7838;

  304.2910  226.2424;

 -969.5442  205.6109;

  296.6569  117.0695;

  520.0000 -200.0000];

by using these waypoints, i get these errors.

presolve, variable S[1,10,1]:

impossible deduced bounds: lower = 520, upper = 30.4558;

difference = 489.544

presolve, variable S[1,10,1]:

impossible deduced bounds: lower = 520, upper = 30.4558;

difference = 489.544

But when I have these as my waypoints,

V = [-5 -2;

 3  4;

  1 6;

 -2  5;

  4  3;

 -5 -2];

it runs normally.

Here is how I declared my model file parameters.

param n{N};                                         #size of waypts for vehicle k

set N ordered;                                      #vehicle numbering

param T integer >=1;                                #maximal horizon

var b{k in N, w in 1..n[k], i in 0..T} binary;      #visit variable  

Is there a way to fix that error for feasibility?

Thank you 

[AMPL Google Group]

You haven't shown how you have declared S in the model file. The error message is telling you that you have set lower bound of S to 520 and upper bound to 30.4558. It's not possible to have a variable greater than 520, but lower than 30.4558. I would recommend you to check the upper and lower bound of S variable. If you face any problem in understanding the bounds, you can reply here with your model and data file. That will help us to understand how you declared S.

Thanks,


--
Dr. Paras Tiwari
am...@googlegroups.com

{#HS:631046433-16798#}

--
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ampl+uns...@googlegroups.com.
To post to this group, send email to am...@googlegroups.com.
Visit this group at https://groups.google.com/group/ampl.
For more options, visit https://groups.google.com/d/optout.

[AMPL Google Group]

Here is more information about the problem:

A message like

presolve, variable S[1,10,1]:
impossible deduced bounds: lower = 520, upper = 30.4558;
difference = 489.544

indicates that AMPL's presolve phase has determined impossible bounds on the specified variable. Presolve performs multiple passes that use information in the constraints to tighten the bounds on some of the variables, and if it gets to a point where a variable's tightened lower bound is higher than the corresponding tightened upper bound, it displays this message. Whenever this message is seen, it can be concluded that there is no feasible solution to the constraints.

Because this is a message from presolve, it does not imply that there is anything wrong with the bounds specified in the model. The user needs to use his knowledge of the model and data to figure out why the constraints imply a lower bound of 520 and an upper bound of 30.4558 for S[1,10,1] -- after which the model or data can be fixed. If the problem is linear then it is possible to get some help by setting "option presolve 0;" and using the IIS-finding feature of one of the MIP solvers.

--
Dr. Paras Tiwari
am...@googlegroups.com

{#HS:631046433-16798#}

On Tue, Jul 31, 2018 at 3:31 AM UTC, AMPL Google Group <am...@googlegroups.com> wrote:

You haven't shown how you have declared S in the model file. The error message is telling you that you have set lower bound of S to 520 and upper bound to 30.4558. It's not possible to have a variable greater than 520, but lower than 30.4558. I would recommend you to check the upper and lower bound of S variable. If you face any problem in understanding the bounds, you can reply here with your model and data file. That will help us to understand how you declared S.

Thanks,


--
Dr. Paras Tiwari
am...@googlegroups.com

On Mon, Jul 30, 2018 at 5:35 PM UTC, Ampl Modeling Language <am...@googlegroups.com> wrote:

[vinhk...@gmail.com]

Thank you all for helping. I will further look into it.