# Impossible deduced bounds

Hi, I am new to AMPL. I was changing the original mod file to add a few different things.

I get the following error:

``````presolve, variable DY[1,4]:
impossible deduced bounds: lower = 11.5142, upper = 8.48579;
difference = 3.02843
presolve, variable DX[1,4]:
impossible deduced bounds: lower = -21.9858, upper = -25.0142;
difference = 3.02843
presolve, variable DY[0,4]:
impossible deduced bounds: lower = 11.5142, upper = 8.48579;
difference = 3.02843
presolve, variable DX[0,4]:
impossible deduced bounds: lower = -21.9858, upper = -25.0142;
difference = 3.02843
presolve, variable CY[1,4]:
impossible deduced bounds: lower = 11.5142, upper = 8.48579;
difference = 3.02843
3979 presolve messages suppressed.
``````

The original mod file part:

``````## Collison avoidance ##

s.t. eq_PPPoutsideABCD {i in {1..Nfe}, pp in {1..Nv}, nn in {1..Nobs}, jj in {1..4}}:
abs((AX[i,pp] - PPP[nn,jj,1])*(BY[i,pp] - PPP[nn,jj,2]) - (AY[i,pp] - PPP[nn,jj,2])*(BX[i,pp] - PPP[nn,jj,1])) * 0.5 + abs((BX[i,pp] - PPP[nn,jj,1])*(CY[i,pp] - PPP[nn,jj,2]) - (BY[i,pp] - PPP[nn,jj,2])*(CX[i,pp] - PPP[nn,jj,1])) * 0.5 + abs((CX[i,pp] - PPP[nn,jj,1])*(DY[i,pp] - PPP[nn,jj,2]) - (CY[i,pp] - PPP[nn,jj,2])*(DX[i,pp] - PPP[nn,jj,1])) * 0.5 + abs((DX[i,pp] - PPP[nn,jj,1])*(AY[i,pp] - PPP[nn,jj,2]) - (DY[i,pp] - PPP[nn,jj,2])*(AX[i,pp] - PPP[nn,jj,1])) * 0.5 >= 4 + 0.1;

s.t. eq_AoutsidePRECTANGLEPPP {i in {1..Nfe}, pp in {1..Nv}, nn in {1..Nobs}}:
abs((PPP[nn,1,1] - AX[i,pp])*( PPP[nn,2,2] - AY[i,pp]) - (PPP[nn,1,2] - AY[i,pp])*(PPP[nn,2,1] - AX[i,pp])) * 0.5 + abs((PPP[nn,2,1] - AX[i,pp])*(PPP[nn,3,2] - AY[i,pp]) - (PPP[nn,2,2] - AY[i,pp])*( PPP[nn,3,1] - AX[i,pp])) * 0.5 + abs((PPP[nn,3,1] - AX[i,pp])*( PPP[nn,4,2] - AY[i,pp]) - (PPP[nn,3,2] - AY[i,pp])*( PPP[nn,4,1] - AX[i,pp])) * 0.5 + abs((PPP[nn,4,1] - AX[i,pp])*( PPP[nn,1,2] - AY[i,pp]) - (PPP[nn,4,2] - AY[i,pp])*( PPP[nn,1,1] - AX[i,pp])) * 0.5 >= AREA[nn] + 0.1;

s.t. eq_BoutsidePRECTANGLEPPP {i in {1..Nfe}, pp in {1..Nv}, nn in {1..Nobs}}:
abs((PPP[nn,1,1] - BX[i,pp])*( PPP[nn,2,2] - BY[i,pp]) - (PPP[nn,1,2] - BY[i,pp])*(PPP[nn,2,1] - BX[i,pp])) * 0.5 + abs((PPP[nn,2,1] - BX[i,pp])*(PPP[nn,3,2] - BY[i,pp]) - (PPP[nn,2,2] - BY[i,pp])*( PPP[nn,3,1] - BX[i,pp])) * 0.5 + abs((PPP[nn,3,1] - BX[i,pp])*( PPP[nn,4,2] - BY[i,pp]) - (PPP[nn,3,2] - BY[i,pp])*( PPP[nn,4,1] - BX[i,pp])) * 0.5 + abs((PPP[nn,4,1] - BX[i,pp])*( PPP[nn,1,2] - BY[i,pp]) - (PPP[nn,4,2] - BY[i,pp])*( PPP[nn,1,1] - BX[i,pp])) * 0.5 >= AREA[nn] + 0.1;

s.t. eq_CoutsidePRECTANGLEPPP {i in {1..Nfe}, pp in {1..Nv}, nn in {1..Nobs}}:
abs((PPP[nn,1,1] - CX[i,pp])*( PPP[nn,2,2] - CY[i,pp]) - (PPP[nn,1,2] - CY[i,pp])*(PPP[nn,2,1] - CX[i,pp])) * 0.5 + abs((PPP[nn,2,1] - CX[i,pp])*(PPP[nn,3,2] - CY[i,pp]) - (PPP[nn,2,2] - CY[i,pp])*( PPP[nn,3,1] - CX[i,pp])) * 0.5 + abs((PPP[nn,3,1] - CX[i,pp])*( PPP[nn,4,2] - CY[i,pp]) - (PPP[nn,3,2] - CY[i,pp])*( PPP[nn,4,1] - CX[i,pp])) * 0.5 + abs((PPP[nn,4,1] - CX[i,pp])*( PPP[nn,1,2] - CY[i,pp]) - (PPP[nn,4,2] - CY[i,pp])*( PPP[nn,1,1] - CX[i,pp])) * 0.5 >= AREA[nn] + 0.1;

s.t. eq_DoutsidePRECTANGLEPPP {i in {1..Nfe}, pp in {1..Nv}, nn in {1..Nobs}}:
abs((PPP[nn,1,1] - DX[i,pp])*( PPP[nn,2,2] - DY[i,pp]) - (PPP[nn,1,2] - DY[i,pp])*(PPP[nn,2,1] - DX[i,pp])) * 0.5 + abs((PPP[nn,2,1] - DX[i,pp])*(PPP[nn,3,2] - DY[i,pp]) - (PPP[nn,2,2] - DY[i,pp])*( PPP[nn,3,1] - DX[i,pp])) * 0.5 + abs((PPP[nn,3,1] - DX[i,pp])*( PPP[nn,4,2] - DY[i,pp]) - (PPP[nn,3,2] - DY[i,pp])*( PPP[nn,4,1] - DX[i,pp])) * 0.5 + abs((PPP[nn,4,1] - DX[i,pp])*( PPP[nn,1,2] - DY[i,pp]) - (PPP[nn,4,2] - DY[i,pp])*( PPP[nn,1,1] - DX[i,pp])) * 0.5 >= AREA[nn] + 0.1;
``````

I changed it to

Any help on this would be appreciated? Thanks.

The error messages are telling you that your problem has no feasible solution: there is no way to assign values to the variables that are within the variables’ bounds and that also satisfy all of the constraints. As an example, consider the first message:

``````presolve, variable DY[1,4]:
impossible deduced bounds: lower = 11.5142, upper = 8.48579;
``````

Based on an analysis of the constraints where variable `DY[1,4]` appears, AMPL has proved that it must be >= 11.5142 in any feasible solution (“lower = 11.5142”), and has also proved that it must be <= 8.48579 in any feasible solution (“upper = 8.48579”). Since a variable cannot be both >= 11.5142 and also <= 8.48579, this proves that no feasible solution is possible.

This proof of infeasibility occurs in AMPL’s “presolve” phase, while the problem is being generated. Thus the problem is never actually sent to the solver.

The infeasibility may be caused by an error in your model. Or, your data may not permit a feasible solution; for example, some capacities may be too low to permit all demands to be met. It is not possible to say more about the cause, however, based only on the information that you have given. To get more help, you should upload your files (using the up-arrow icon of the forum editor).

Thank you for your response. I found that there was an issue with the data in the model, and after changing the symbols, the problem no longer occurred.