search

MOSEK / Mosek.jl

Interface to the Mosek solver in Julia
MIT License
49 stars 28 forks source link

MOI.OPTIMAL status but infeasible solution #212

Closed SobhanMP closed 2 years ago

SobhanMP commented 2 years ago 
g = [1, 2, 3]
model = Model(Mosek.Optimizer)

@variables(model, begin
        p[eachindex(g)] >= 0
        Ω[eachindex(g)]
end)

@constraint(model, sum(p) == 1)
@constraint(model, [i=eachindex(g)], [1, p[i], Ω[i]] in MOI.ExponentialCone())
@objective(model, Max, sum(p .* g) + 0.0 * sum(Ω))
latex_formulation(model)

image

In this model (that was wrong it should've been [Ω[i], p[i], 1] in MOI.ExponentialCone()) but regardless, when optimizing the model reaches optimality but the constraint sum(p) == 1 is not satisfied.

optimize!(model)
Problem
  Name                   :                 
  Objective sense        : max             
  Type                   : CONIC (conic optimization problem)
  Constraints            : 10              
  Cones                  : 3               
  Scalar variables       : 15              
  Matrix variables       : 0               
  Integer variables      : 0               

Optimizer started.
Presolve started.
Linear dependency checker started.
Linear dependency checker terminated.
Eliminator started.
Freed constraints in eliminator : 0
Eliminator terminated.
Eliminator - tries                  : 1                 time                   : 0.00            
Lin. dep.  - tries                  : 1                 time                   : 0.00            
Lin. dep.  - number                 : 0               
Presolve terminated. Time: 0.00    
Problem
  Name                   :                 
  Objective sense        : max             
  Type                   : CONIC (conic optimization problem)
  Constraints            : 10              
  Cones                  : 3               
  Scalar variables       : 15              
  Matrix variables       : 0               
  Integer variables      : 0               

Optimizer  - threads                : 18              
Optimizer  - solved problem         : the primal      
Optimizer  - Constraints            : 1
Optimizer  - Cones                  : 3
Optimizer  - Scalar variables       : 9                 conic                  : 9               
Optimizer  - Semi-definite variables: 0                 scalarized             : 0               
Factor     - setup time             : 0.00              dense det. time        : 0.00            
Factor     - ML order time          : 0.00              GP order time          : 0.00            
Factor     - nonzeros before factor : 1                 after factor           : 1               
Factor     - dense dim.             : 0                 flops                  : 1.30e+01        
ITE PFEAS    DFEAS    GFEAS    PRSTATUS   POBJ              DOBJ              MU       TIME  
0   1.8e+00  3.8e+00  6.3e+00  0.00e+00   4.830612010e+00   -2.483515197e+00  1.0e+00  0.00  
1   2.1e-01  4.4e-01  7.8e-01  -3.14e-01  3.018859342e+00   1.785519550e-01   1.2e-01  0.00  
2   3.2e-02  6.6e-02  9.7e-02  1.22e-01   2.222927263e+00   9.727063870e-01   1.7e-02  0.00  
3   6.4e-03  1.3e-02  4.2e-03  9.07e-01   2.117075200e+00   1.993968819e+00   3.5e-03  0.00  
4   1.0e-03  2.2e-03  4.9e-04  6.46e-01   2.270224493e+00   2.231262552e+00   5.7e-04  0.00  
5   2.2e-04  4.7e-04  1.1e-04  1.72e-01   2.412201194e+00   2.383051296e+00   1.2e-04  0.00  
6   4.7e-05  9.7e-05  1.8e-05  3.65e-01   2.475835861e+00   2.459845099e+00   2.5e-05  0.00  
7   1.7e-05  3.4e-05  7.4e-06  5.85e-02   2.541615728e+00   2.522650249e+00   9.0e-06  0.00  
8   3.4e-06  7.1e-06  1.1e-06  4.22e-01   2.574995923e+00   2.566225740e+00   1.9e-06  0.00  
9   1.3e-06  2.7e-06  4.7e-07  9.78e-02   2.619258121e+00   2.607690011e+00   7.1e-07  0.00  
10  2.9e-07  6.0e-07  7.2e-08  4.67e-01   2.637973135e+00   2.632572432e+00   1.6e-07  0.00  
11  1.1e-07  2.4e-07  3.5e-08  9.35e-02   2.671000491e+00   2.663023181e+00   6.2e-08  0.00  
12  2.4e-08  5.1e-08  5.0e-09  4.87e-01   2.684172907e+00   2.680637816e+00   1.3e-08  0.00  
13  8.1e-09  1.7e-08  2.1e-09  6.02e-02   2.713815830e+00   2.707895975e+00   4.4e-09  0.00  
14  1.9e-09  3.9e-09  3.2e-10  5.33e-01   2.721289042e+00   2.718900850e+00   1.0e-09  0.00  
15  6.4e-10  1.3e-09  1.4e-10  4.71e-02   2.743921987e+00   2.739772216e+00   3.5e-10  0.00  
16  1.4e-10  2.9e-10  2.0e-11  4.96e-01   2.751292501e+00   2.749514064e+00   7.5e-11  0.00  
17  3.9e-11  8.2e-11  7.8e-12  2.26e-02   2.771969917e+00   2.768655197e+00   2.2e-11  0.00  
18  1.0e-11  2.1e-11  1.2e-12  5.96e-01   2.775657313e+00   2.774381591e+00   5.5e-12  0.00  
19  3.5e-12  7.2e-12  5.8e-13  3.15e-02   2.789987283e+00   2.787650836e+00   1.9e-12  0.00  
20  7.8e-13  1.6e-12  8.9e-14  4.81e-01   2.795357759e+00   2.794265755e+00   4.2e-13  0.00  
21  2.4e-13  5.0e-13  3.4e-14  2.99e-02   2.805805254e+00   2.804135451e+00   1.3e-13  0.00  
22  4.8e-14  1.0e-13  5.7e-15  2.71e-01   2.813539824e+00   2.812365436e+00   2.6e-14  0.00  
23  1.5e-14  3.1e-14  2.2e-15  -2.40e-03  2.822450247e+00   2.820697208e+00   8.3e-15  0.00  
24  3.5e-15  7.4e-15  3.8e-16  3.85e-01   2.827082078e+00   2.826125811e+00   1.9e-15  0.00  
25  9.5e-16  2.0e-15  1.3e-16  -4.85e-02  2.836327732e+00   2.834682575e+00   5.2e-16  0.00  
26  2.2e-16  4.6e-16  2.4e-17  3.36e-01   2.840379465e+00   2.839404879e+00   1.2e-16  0.00  
27  7.3e-17  1.5e-16  9.7e-18  -3.25e-02  2.846980326e+00   2.845491145e+00   4.0e-17  0.00  
28  1.8e-17  3.6e-17  1.7e-18  3.83e-01   2.850175560e+00   2.849408256e+00   9.6e-18  0.00  
29  5.5e-18  1.1e-17  6.6e-19  -3.01e-02  2.856158109e+00   2.854939799e+00   3.0e-18  0.00  
30  1.1e-18  2.4e-18  1.1e-19  2.61e-01   2.860147432e+00   2.859335447e+00   6.2e-19  0.00  
31  4.0e-19  8.4e-19  4.8e-20  9.84e-03   2.864801063e+00   2.863626950e+00   2.2e-19  0.00  
32  1.0e-19  2.1e-19  8.5e-21  4.07e-01   2.867208551e+00   2.866604890e+00   5.5e-20  0.00  
33  3.1e-20  6.3e-20  3.4e-21  -5.87e-02  2.872464596e+00   2.871410124e+00   1.7e-20  0.00  
34  7.1e-21  1.5e-20  5.9e-22  3.37e-01   2.874830484e+00   2.874233898e+00   3.8e-21  0.00  
35  2.4e-21  4.9e-21  2.5e-22  -3.94e-02  2.878902315e+00   2.877980792e+00   1.3e-21  0.00  
36  6.1e-22  1.2e-21  4.4e-23  3.81e-01   2.880864248e+00   2.880386453e+00   3.2e-22  0.00  
37  2.3e-22  3.8e-22  1.8e-23  -3.92e-02  2.884610205e+00   2.883845993e+00   1.0e-22  0.00  
38  4.8e-23  7.9e-23  3.0e-24  2.45e-01   2.887246096e+00   2.886726479e+00   2.1e-23  0.00  
39  9.3e-23  2.8e-23  1.3e-24  6.01e-03   2.890275629e+00   2.889520967e+00   7.3e-24  0.00  
40  4.1e-23  7.0e-24  2.3e-25  4.04e-01   2.891854464e+00   2.891463497e+00   1.8e-24  0.00  
41  1.7e-23  2.0e-24  9.0e-26  -7.17e-02  2.895496116e+00   2.894787867e+00   5.3e-25  0.00  
42  6.4e-24  4.7e-25  1.5e-26  3.56e-01   2.896958288e+00   2.896575882e+00   1.3e-25  0.00  
43  1.1e-23  1.6e-25  6.6e-27  -4.71e-02  2.899796396e+00   2.899186023e+00   4.2e-26  0.00  
44  2.6e-24  4.2e-26  1.1e-27  3.80e-01   2.901144025e+00   2.900826139e+00   1.0e-26  0.00  
45  1.2e-24  1.5e-26  4.8e-28  -3.93e-02  2.903739795e+00   2.903228499e+00   3.2e-27  0.00  
46  3.0e-25  8.5e-28  4.5e-29  2.39e-01   2.905594920e+00   2.905240326e+00   6.6e-28  0.00  
47  8.9e-26  4.0e-28  6.0e-30  -3.37e-03  2.907754518e+00   2.907232163e+00   2.3e-28  0.00  
Optimizer terminated. Time: 0.00

the solution found is

[2.3028378190668097e-14, 2.402867139391792e-14, 7.127137945447063e-13]

which violates the sum(p) == 1 constraint.

Is this a bug? The correct problem works fine BTW. i'm using mosek 9.3.18, mosek.jl 1.2.1 adn, mosek tools 0.12.0 on julia 1.6

erling-d-andersen commented 2 years ago

I very much doubt Mosek says it found an optimal solution. I think it said illposed.

Since the Mosek solution summary is not shown I cannot say for sure.

lør. 19. mar. 2022 21.28 skrev Sobhan Mohammadpour @.***

:

g = [1, 2, 3]

model = Model(Mosek.Optimizer)

@variables(model, begin

    p[eachindex(g)] >= 0

    Ω[eachindex(g)]

end)

@constraint(model, sum(p) == 1) @constraint(model, [i=eachindex(g)], [1, p[i], Ω[i]] in MOI.ExponentialCone()) @objective(model, Max, sum(p . g) + 0.0 sum(Ω)) latex_formulation(model)

[image: image] https://user-images.githubusercontent.com/2301159/159137223-ae1a55c9-f42e-4d11-b35b-a8dd01749d03.png

In this model (that was wrong it should've been [Ω[i], p[i], 1] in MOI.ExponentialCone()) but regardless, when optimizing the model reaches optimality but the constraint sum(p) == 1 is not satisfied.

optimize!(model)

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer started.

Presolve started.

Linear dependency checker started.

Linear dependency checker terminated.

Eliminator started.

Freed constraints in eliminator : 0

Eliminator terminated.

Eliminator - tries : 1 time : 0.00

Lin. dep. - tries : 1 time : 0.00

Lin. dep. - number : 0

Presolve terminated. Time: 0.00

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer - threads : 18

Optimizer - solved problem : the primal

Optimizer - Constraints : 1

Optimizer - Cones : 3

Optimizer - Scalar variables : 9 conic : 9

Optimizer - Semi-definite variables: 0 scalarized : 0

Factor - setup time : 0.00 dense det. time : 0.00

Factor - ML order time : 0.00 GP order time : 0.00

Factor - nonzeros before factor : 1 after factor : 1

Factor - dense dim. : 0 flops : 1.30e+01

ITE PFEAS DFEAS GFEAS PRSTATUS POBJ DOBJ MU TIME

0 1.8e+00 3.8e+00 6.3e+00 0.00e+00 4.830612010e+00 -2.483515197e+00 1.0e+00 0.00

1 2.1e-01 4.4e-01 7.8e-01 -3.14e-01 3.018859342e+00 1.785519550e-01 1.2e-01 0.00

2 3.2e-02 6.6e-02 9.7e-02 1.22e-01 2.222927263e+00 9.727063870e-01 1.7e-02 0.00

3 6.4e-03 1.3e-02 4.2e-03 9.07e-01 2.117075200e+00 1.993968819e+00 3.5e-03 0.00

4 1.0e-03 2.2e-03 4.9e-04 6.46e-01 2.270224493e+00 2.231262552e+00 5.7e-04 0.00

5 2.2e-04 4.7e-04 1.1e-04 1.72e-01 2.412201194e+00 2.383051296e+00 1.2e-04 0.00

6 4.7e-05 9.7e-05 1.8e-05 3.65e-01 2.475835861e+00 2.459845099e+00 2.5e-05 0.00

7 1.7e-05 3.4e-05 7.4e-06 5.85e-02 2.541615728e+00 2.522650249e+00 9.0e-06 0.00

8 3.4e-06 7.1e-06 1.1e-06 4.22e-01 2.574995923e+00 2.566225740e+00 1.9e-06 0.00

9 1.3e-06 2.7e-06 4.7e-07 9.78e-02 2.619258121e+00 2.607690011e+00 7.1e-07 0.00

10 2.9e-07 6.0e-07 7.2e-08 4.67e-01 2.637973135e+00 2.632572432e+00 1.6e-07 0.00

11 1.1e-07 2.4e-07 3.5e-08 9.35e-02 2.671000491e+00 2.663023181e+00 6.2e-08 0.00

12 2.4e-08 5.1e-08 5.0e-09 4.87e-01 2.684172907e+00 2.680637816e+00 1.3e-08 0.00

13 8.1e-09 1.7e-08 2.1e-09 6.02e-02 2.713815830e+00 2.707895975e+00 4.4e-09 0.00

14 1.9e-09 3.9e-09 3.2e-10 5.33e-01 2.721289042e+00 2.718900850e+00 1.0e-09 0.00

15 6.4e-10 1.3e-09 1.4e-10 4.71e-02 2.743921987e+00 2.739772216e+00 3.5e-10 0.00

16 1.4e-10 2.9e-10 2.0e-11 4.96e-01 2.751292501e+00 2.749514064e+00 7.5e-11 0.00

17 3.9e-11 8.2e-11 7.8e-12 2.26e-02 2.771969917e+00 2.768655197e+00 2.2e-11 0.00

18 1.0e-11 2.1e-11 1.2e-12 5.96e-01 2.775657313e+00 2.774381591e+00 5.5e-12 0.00

19 3.5e-12 7.2e-12 5.8e-13 3.15e-02 2.789987283e+00 2.787650836e+00 1.9e-12 0.00

20 7.8e-13 1.6e-12 8.9e-14 4.81e-01 2.795357759e+00 2.794265755e+00 4.2e-13 0.00

21 2.4e-13 5.0e-13 3.4e-14 2.99e-02 2.805805254e+00 2.804135451e+00 1.3e-13 0.00

22 4.8e-14 1.0e-13 5.7e-15 2.71e-01 2.813539824e+00 2.812365436e+00 2.6e-14 0.00

23 1.5e-14 3.1e-14 2.2e-15 -2.40e-03 2.822450247e+00 2.820697208e+00 8.3e-15 0.00

24 3.5e-15 7.4e-15 3.8e-16 3.85e-01 2.827082078e+00 2.826125811e+00 1.9e-15 0.00

25 9.5e-16 2.0e-15 1.3e-16 -4.85e-02 2.836327732e+00 2.834682575e+00 5.2e-16 0.00

26 2.2e-16 4.6e-16 2.4e-17 3.36e-01 2.840379465e+00 2.839404879e+00 1.2e-16 0.00

27 7.3e-17 1.5e-16 9.7e-18 -3.25e-02 2.846980326e+00 2.845491145e+00 4.0e-17 0.00

28 1.8e-17 3.6e-17 1.7e-18 3.83e-01 2.850175560e+00 2.849408256e+00 9.6e-18 0.00

29 5.5e-18 1.1e-17 6.6e-19 -3.01e-02 2.856158109e+00 2.854939799e+00 3.0e-18 0.00

30 1.1e-18 2.4e-18 1.1e-19 2.61e-01 2.860147432e+00 2.859335447e+00 6.2e-19 0.00

31 4.0e-19 8.4e-19 4.8e-20 9.84e-03 2.864801063e+00 2.863626950e+00 2.2e-19 0.00

32 1.0e-19 2.1e-19 8.5e-21 4.07e-01 2.867208551e+00 2.866604890e+00 5.5e-20 0.00

33 3.1e-20 6.3e-20 3.4e-21 -5.87e-02 2.872464596e+00 2.871410124e+00 1.7e-20 0.00

34 7.1e-21 1.5e-20 5.9e-22 3.37e-01 2.874830484e+00 2.874233898e+00 3.8e-21 0.00

35 2.4e-21 4.9e-21 2.5e-22 -3.94e-02 2.878902315e+00 2.877980792e+00 1.3e-21 0.00

36 6.1e-22 1.2e-21 4.4e-23 3.81e-01 2.880864248e+00 2.880386453e+00 3.2e-22 0.00

37 2.3e-22 3.8e-22 1.8e-23 -3.92e-02 2.884610205e+00 2.883845993e+00 1.0e-22 0.00

38 4.8e-23 7.9e-23 3.0e-24 2.45e-01 2.887246096e+00 2.886726479e+00 2.1e-23 0.00

39 9.3e-23 2.8e-23 1.3e-24 6.01e-03 2.890275629e+00 2.889520967e+00 7.3e-24 0.00

40 4.1e-23 7.0e-24 2.3e-25 4.04e-01 2.891854464e+00 2.891463497e+00 1.8e-24 0.00

41 1.7e-23 2.0e-24 9.0e-26 -7.17e-02 2.895496116e+00 2.894787867e+00 5.3e-25 0.00

42 6.4e-24 4.7e-25 1.5e-26 3.56e-01 2.896958288e+00 2.896575882e+00 1.3e-25 0.00

43 1.1e-23 1.6e-25 6.6e-27 -4.71e-02 2.899796396e+00 2.899186023e+00 4.2e-26 0.00

44 2.6e-24 4.2e-26 1.1e-27 3.80e-01 2.901144025e+00 2.900826139e+00 1.0e-26 0.00

45 1.2e-24 1.5e-26 4.8e-28 -3.93e-02 2.903739795e+00 2.903228499e+00 3.2e-27 0.00

46 3.0e-25 8.5e-28 4.5e-29 2.39e-01 2.905594920e+00 2.905240326e+00 6.6e-28 0.00

47 8.9e-26 4.0e-28 6.0e-30 -3.37e-03 2.907754518e+00 2.907232163e+00 2.3e-28 0.00

Optimizer terminated. Time: 0.00

the solution found is

[2.3028378190668097e-14, 2.402867139391792e-14, 7.127137945447063e-13]

which violates the sum(p) == 1 constraint.

Is this a mosek bug? i'm using mosek 9.3.18, mosek.jl 1.2.1 adn, mosek tools 0.12.0 on julia 1.6

— Reply to this email directly, view it on GitHub https://github.com/MOSEK/Mosek.jl/issues/212, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACOAE4K722BYO6R3YVX5W63VAY2F3ANCNFSM5REPCP4A . You are receiving this because you are subscribed to this thread.Message ID: @.***>

erling-d-andersen commented 2 years ago

If you want us to look at the problem you can dump it to a file using the instructions

https://docs.mosek.com/latest/faq/faq.html

and email it to

@.***

Den søn. 20. mar. 2022 kl. 16.15 skrev Erling Andersen < @.***>:

I very much doubt Mosek says it found an optimal solution. I think it said illposed.

Since the Mosek solution summary is not shown I cannot say for sure.

lør. 19. mar. 2022 21.28 skrev Sobhan Mohammadpour < @.***>:

g = [1, 2, 3]

model = Model(Mosek.Optimizer)

@variables(model, begin

    p[eachindex(g)] >= 0

    Ω[eachindex(g)]

end)

@constraint(model, sum(p) == 1) @constraint(model, [i=eachindex(g)], [1, p[i], Ω[i]] in MOI.ExponentialCone()) @objective(model, Max, sum(p . g) + 0.0 sum(Ω)) latex_formulation(model)

[image: image] https://user-images.githubusercontent.com/2301159/159137223-ae1a55c9-f42e-4d11-b35b-a8dd01749d03.png

In this model (that was wrong it should've been [Ω[i], p[i], 1] in MOI.ExponentialCone()) but regardless, when optimizing the model reaches optimality but the constraint sum(p) == 1 is not satisfied.

optimize!(model)

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer started.

Presolve started.

Linear dependency checker started.

Linear dependency checker terminated.

Eliminator started.

Freed constraints in eliminator : 0

Eliminator terminated.

Eliminator - tries : 1 time : 0.00

Lin. dep. - tries : 1 time : 0.00

Lin. dep. - number : 0

Presolve terminated. Time: 0.00

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer - threads : 18

Optimizer - solved problem : the primal

Optimizer - Constraints : 1

Optimizer - Cones : 3

Optimizer - Scalar variables : 9 conic : 9

Optimizer - Semi-definite variables: 0 scalarized : 0

Factor - setup time : 0.00 dense det. time : 0.00

Factor - ML order time : 0.00 GP order time : 0.00

Factor - nonzeros before factor : 1 after factor : 1

Factor - dense dim. : 0 flops : 1.30e+01

ITE PFEAS DFEAS GFEAS PRSTATUS POBJ DOBJ MU TIME

0 1.8e+00 3.8e+00 6.3e+00 0.00e+00 4.830612010e+00 -2.483515197e+00 1.0e+00 0.00

1 2.1e-01 4.4e-01 7.8e-01 -3.14e-01 3.018859342e+00 1.785519550e-01 1.2e-01 0.00

2 3.2e-02 6.6e-02 9.7e-02 1.22e-01 2.222927263e+00 9.727063870e-01 1.7e-02 0.00

3 6.4e-03 1.3e-02 4.2e-03 9.07e-01 2.117075200e+00 1.993968819e+00 3.5e-03 0.00

4 1.0e-03 2.2e-03 4.9e-04 6.46e-01 2.270224493e+00 2.231262552e+00 5.7e-04 0.00

5 2.2e-04 4.7e-04 1.1e-04 1.72e-01 2.412201194e+00 2.383051296e+00 1.2e-04 0.00

6 4.7e-05 9.7e-05 1.8e-05 3.65e-01 2.475835861e+00 2.459845099e+00 2.5e-05 0.00

7 1.7e-05 3.4e-05 7.4e-06 5.85e-02 2.541615728e+00 2.522650249e+00 9.0e-06 0.00

8 3.4e-06 7.1e-06 1.1e-06 4.22e-01 2.574995923e+00 2.566225740e+00 1.9e-06 0.00

9 1.3e-06 2.7e-06 4.7e-07 9.78e-02 2.619258121e+00 2.607690011e+00 7.1e-07 0.00

10 2.9e-07 6.0e-07 7.2e-08 4.67e-01 2.637973135e+00 2.632572432e+00 1.6e-07 0.00

11 1.1e-07 2.4e-07 3.5e-08 9.35e-02 2.671000491e+00 2.663023181e+00 6.2e-08 0.00

12 2.4e-08 5.1e-08 5.0e-09 4.87e-01 2.684172907e+00 2.680637816e+00 1.3e-08 0.00

13 8.1e-09 1.7e-08 2.1e-09 6.02e-02 2.713815830e+00 2.707895975e+00 4.4e-09 0.00

14 1.9e-09 3.9e-09 3.2e-10 5.33e-01 2.721289042e+00 2.718900850e+00 1.0e-09 0.00

15 6.4e-10 1.3e-09 1.4e-10 4.71e-02 2.743921987e+00 2.739772216e+00 3.5e-10 0.00

16 1.4e-10 2.9e-10 2.0e-11 4.96e-01 2.751292501e+00 2.749514064e+00 7.5e-11 0.00

17 3.9e-11 8.2e-11 7.8e-12 2.26e-02 2.771969917e+00 2.768655197e+00 2.2e-11 0.00

18 1.0e-11 2.1e-11 1.2e-12 5.96e-01 2.775657313e+00 2.774381591e+00 5.5e-12 0.00

19 3.5e-12 7.2e-12 5.8e-13 3.15e-02 2.789987283e+00 2.787650836e+00 1.9e-12 0.00

20 7.8e-13 1.6e-12 8.9e-14 4.81e-01 2.795357759e+00 2.794265755e+00 4.2e-13 0.00

21 2.4e-13 5.0e-13 3.4e-14 2.99e-02 2.805805254e+00 2.804135451e+00 1.3e-13 0.00

22 4.8e-14 1.0e-13 5.7e-15 2.71e-01 2.813539824e+00 2.812365436e+00 2.6e-14 0.00

23 1.5e-14 3.1e-14 2.2e-15 -2.40e-03 2.822450247e+00 2.820697208e+00 8.3e-15 0.00

24 3.5e-15 7.4e-15 3.8e-16 3.85e-01 2.827082078e+00 2.826125811e+00 1.9e-15 0.00

25 9.5e-16 2.0e-15 1.3e-16 -4.85e-02 2.836327732e+00 2.834682575e+00 5.2e-16 0.00

26 2.2e-16 4.6e-16 2.4e-17 3.36e-01 2.840379465e+00 2.839404879e+00 1.2e-16 0.00

27 7.3e-17 1.5e-16 9.7e-18 -3.25e-02 2.846980326e+00 2.845491145e+00 4.0e-17 0.00

28 1.8e-17 3.6e-17 1.7e-18 3.83e-01 2.850175560e+00 2.849408256e+00 9.6e-18 0.00

29 5.5e-18 1.1e-17 6.6e-19 -3.01e-02 2.856158109e+00 2.854939799e+00 3.0e-18 0.00

30 1.1e-18 2.4e-18 1.1e-19 2.61e-01 2.860147432e+00 2.859335447e+00 6.2e-19 0.00

31 4.0e-19 8.4e-19 4.8e-20 9.84e-03 2.864801063e+00 2.863626950e+00 2.2e-19 0.00

32 1.0e-19 2.1e-19 8.5e-21 4.07e-01 2.867208551e+00 2.866604890e+00 5.5e-20 0.00

33 3.1e-20 6.3e-20 3.4e-21 -5.87e-02 2.872464596e+00 2.871410124e+00 1.7e-20 0.00

34 7.1e-21 1.5e-20 5.9e-22 3.37e-01 2.874830484e+00 2.874233898e+00 3.8e-21 0.00

35 2.4e-21 4.9e-21 2.5e-22 -3.94e-02 2.878902315e+00 2.877980792e+00 1.3e-21 0.00

36 6.1e-22 1.2e-21 4.4e-23 3.81e-01 2.880864248e+00 2.880386453e+00 3.2e-22 0.00

37 2.3e-22 3.8e-22 1.8e-23 -3.92e-02 2.884610205e+00 2.883845993e+00 1.0e-22 0.00

38 4.8e-23 7.9e-23 3.0e-24 2.45e-01 2.887246096e+00 2.886726479e+00 2.1e-23 0.00

39 9.3e-23 2.8e-23 1.3e-24 6.01e-03 2.890275629e+00 2.889520967e+00 7.3e-24 0.00

40 4.1e-23 7.0e-24 2.3e-25 4.04e-01 2.891854464e+00 2.891463497e+00 1.8e-24 0.00

41 1.7e-23 2.0e-24 9.0e-26 -7.17e-02 2.895496116e+00 2.894787867e+00 5.3e-25 0.00

42 6.4e-24 4.7e-25 1.5e-26 3.56e-01 2.896958288e+00 2.896575882e+00 1.3e-25 0.00

43 1.1e-23 1.6e-25 6.6e-27 -4.71e-02 2.899796396e+00 2.899186023e+00 4.2e-26 0.00

44 2.6e-24 4.2e-26 1.1e-27 3.80e-01 2.901144025e+00 2.900826139e+00 1.0e-26 0.00

45 1.2e-24 1.5e-26 4.8e-28 -3.93e-02 2.903739795e+00 2.903228499e+00 3.2e-27 0.00

46 3.0e-25 8.5e-28 4.5e-29 2.39e-01 2.905594920e+00 2.905240326e+00 6.6e-28 0.00

47 8.9e-26 4.0e-28 6.0e-30 -3.37e-03 2.907754518e+00 2.907232163e+00 2.3e-28 0.00

Optimizer terminated. Time: 0.00

the solution found is

[2.3028378190668097e-14, 2.402867139391792e-14, 7.127137945447063e-13]

which violates the sum(p) == 1 constraint.

Is this a mosek bug? i'm using mosek 9.3.18, mosek.jl 1.2.1 adn, mosek tools 0.12.0 on julia 1.6

— Reply to this email directly, view it on GitHub https://github.com/MOSEK/Mosek.jl/issues/212, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACOAE4K722BYO6R3YVX5W63VAY2F3ANCNFSM5REPCP4A . You are receiving this because you are subscribed to this thread.Message ID: @.***>

erling-d-andersen commented 2 years ago

When I run Mosek from the command line on the problem

Interior-point solution summary Problem status : ILL_POSED Solution status : DUAL_ILLPOSED_CER Primal. obj: 2.2092271046e-12 nrm: 5e+00 Viol. con: 8e-13 var: 0e+00 cones: 0e+00

So Mosek says dual ill posed which seems correct.

How do you conclude Mosek says the solution is optimal? Do you check a solution status or?

Den man. 21. mar. 2022 kl. 07.02 skrev Erling Andersen < @.***>:

If you want us to look at the problem you can dump it to a file using the instructions

https://docs.mosek.com/latest/faq/faq.html

and email it to

@.***

Den søn. 20. mar. 2022 kl. 16.15 skrev Erling Andersen < @.***>:

I very much doubt Mosek says it found an optimal solution. I think it said illposed.

Since the Mosek solution summary is not shown I cannot say for sure.

lør. 19. mar. 2022 21.28 skrev Sobhan Mohammadpour < @.***>:

g = [1, 2, 3]

model = Model(Mosek.Optimizer)

@variables(model, begin

    p[eachindex(g)] >= 0

    Ω[eachindex(g)]

end)

@constraint(model, sum(p) == 1) @constraint(model, [i=eachindex(g)], [1, p[i], Ω[i]] in MOI.ExponentialCone()) @objective(model, Max, sum(p . g) + 0.0 sum(Ω)) latex_formulation(model)

[image: image] https://user-images.githubusercontent.com/2301159/159137223-ae1a55c9-f42e-4d11-b35b-a8dd01749d03.png

In this model (that was wrong it should've been [Ω[i], p[i], 1] in MOI.ExponentialCone()) but regardless, when optimizing the model reaches optimality but the constraint sum(p) == 1 is not satisfied.

optimize!(model)

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer started.

Presolve started.

Linear dependency checker started.

Linear dependency checker terminated.

Eliminator started.

Freed constraints in eliminator : 0

Eliminator terminated.

Eliminator - tries : 1 time : 0.00

Lin. dep. - tries : 1 time : 0.00

Lin. dep. - number : 0

Presolve terminated. Time: 0.00

Problem

Name :

Objective sense : max

Type : CONIC (conic optimization problem)

Constraints : 10

Cones : 3

Scalar variables : 15

Matrix variables : 0

Integer variables : 0

Optimizer - threads : 18

Optimizer - solved problem : the primal

Optimizer - Constraints : 1

Optimizer - Cones : 3

Optimizer - Scalar variables : 9 conic : 9

Optimizer - Semi-definite variables: 0 scalarized : 0

Factor - setup time : 0.00 dense det. time : 0.00

Factor - ML order time : 0.00 GP order time : 0.00

Factor - nonzeros before factor : 1 after factor : 1

Factor - dense dim. : 0 flops : 1.30e+01

ITE PFEAS DFEAS GFEAS PRSTATUS POBJ DOBJ MU TIME

0 1.8e+00 3.8e+00 6.3e+00 0.00e+00 4.830612010e+00 -2.483515197e+00 1.0e+00 0.00

1 2.1e-01 4.4e-01 7.8e-01 -3.14e-01 3.018859342e+00 1.785519550e-01 1.2e-01 0.00

2 3.2e-02 6.6e-02 9.7e-02 1.22e-01 2.222927263e+00 9.727063870e-01 1.7e-02 0.00

3 6.4e-03 1.3e-02 4.2e-03 9.07e-01 2.117075200e+00 1.993968819e+00 3.5e-03 0.00

4 1.0e-03 2.2e-03 4.9e-04 6.46e-01 2.270224493e+00 2.231262552e+00 5.7e-04 0.00

5 2.2e-04 4.7e-04 1.1e-04 1.72e-01 2.412201194e+00 2.383051296e+00 1.2e-04 0.00

6 4.7e-05 9.7e-05 1.8e-05 3.65e-01 2.475835861e+00 2.459845099e+00 2.5e-05 0.00

7 1.7e-05 3.4e-05 7.4e-06 5.85e-02 2.541615728e+00 2.522650249e+00 9.0e-06 0.00

8 3.4e-06 7.1e-06 1.1e-06 4.22e-01 2.574995923e+00 2.566225740e+00 1.9e-06 0.00

9 1.3e-06 2.7e-06 4.7e-07 9.78e-02 2.619258121e+00 2.607690011e+00 7.1e-07 0.00

10 2.9e-07 6.0e-07 7.2e-08 4.67e-01 2.637973135e+00 2.632572432e+00 1.6e-07 0.00

11 1.1e-07 2.4e-07 3.5e-08 9.35e-02 2.671000491e+00 2.663023181e+00 6.2e-08 0.00

12 2.4e-08 5.1e-08 5.0e-09 4.87e-01 2.684172907e+00 2.680637816e+00 1.3e-08 0.00

13 8.1e-09 1.7e-08 2.1e-09 6.02e-02 2.713815830e+00 2.707895975e+00 4.4e-09 0.00

14 1.9e-09 3.9e-09 3.2e-10 5.33e-01 2.721289042e+00 2.718900850e+00 1.0e-09 0.00

15 6.4e-10 1.3e-09 1.4e-10 4.71e-02 2.743921987e+00 2.739772216e+00 3.5e-10 0.00

16 1.4e-10 2.9e-10 2.0e-11 4.96e-01 2.751292501e+00 2.749514064e+00 7.5e-11 0.00

17 3.9e-11 8.2e-11 7.8e-12 2.26e-02 2.771969917e+00 2.768655197e+00 2.2e-11 0.00

18 1.0e-11 2.1e-11 1.2e-12 5.96e-01 2.775657313e+00 2.774381591e+00 5.5e-12 0.00

19 3.5e-12 7.2e-12 5.8e-13 3.15e-02 2.789987283e+00 2.787650836e+00 1.9e-12 0.00

20 7.8e-13 1.6e-12 8.9e-14 4.81e-01 2.795357759e+00 2.794265755e+00 4.2e-13 0.00

21 2.4e-13 5.0e-13 3.4e-14 2.99e-02 2.805805254e+00 2.804135451e+00 1.3e-13 0.00

22 4.8e-14 1.0e-13 5.7e-15 2.71e-01 2.813539824e+00 2.812365436e+00 2.6e-14 0.00

23 1.5e-14 3.1e-14 2.2e-15 -2.40e-03 2.822450247e+00 2.820697208e+00 8.3e-15 0.00

24 3.5e-15 7.4e-15 3.8e-16 3.85e-01 2.827082078e+00 2.826125811e+00 1.9e-15 0.00

25 9.5e-16 2.0e-15 1.3e-16 -4.85e-02 2.836327732e+00 2.834682575e+00 5.2e-16 0.00

26 2.2e-16 4.6e-16 2.4e-17 3.36e-01 2.840379465e+00 2.839404879e+00 1.2e-16 0.00

27 7.3e-17 1.5e-16 9.7e-18 -3.25e-02 2.846980326e+00 2.845491145e+00 4.0e-17 0.00

28 1.8e-17 3.6e-17 1.7e-18 3.83e-01 2.850175560e+00 2.849408256e+00 9.6e-18 0.00

29 5.5e-18 1.1e-17 6.6e-19 -3.01e-02 2.856158109e+00 2.854939799e+00 3.0e-18 0.00

30 1.1e-18 2.4e-18 1.1e-19 2.61e-01 2.860147432e+00 2.859335447e+00 6.2e-19 0.00

31 4.0e-19 8.4e-19 4.8e-20 9.84e-03 2.864801063e+00 2.863626950e+00 2.2e-19 0.00

32 1.0e-19 2.1e-19 8.5e-21 4.07e-01 2.867208551e+00 2.866604890e+00 5.5e-20 0.00

33 3.1e-20 6.3e-20 3.4e-21 -5.87e-02 2.872464596e+00 2.871410124e+00 1.7e-20 0.00

34 7.1e-21 1.5e-20 5.9e-22 3.37e-01 2.874830484e+00 2.874233898e+00 3.8e-21 0.00

35 2.4e-21 4.9e-21 2.5e-22 -3.94e-02 2.878902315e+00 2.877980792e+00 1.3e-21 0.00

36 6.1e-22 1.2e-21 4.4e-23 3.81e-01 2.880864248e+00 2.880386453e+00 3.2e-22 0.00

37 2.3e-22 3.8e-22 1.8e-23 -3.92e-02 2.884610205e+00 2.883845993e+00 1.0e-22 0.00

38 4.8e-23 7.9e-23 3.0e-24 2.45e-01 2.887246096e+00 2.886726479e+00 2.1e-23 0.00

39 9.3e-23 2.8e-23 1.3e-24 6.01e-03 2.890275629e+00 2.889520967e+00 7.3e-24 0.00

40 4.1e-23 7.0e-24 2.3e-25 4.04e-01 2.891854464e+00 2.891463497e+00 1.8e-24 0.00

41 1.7e-23 2.0e-24 9.0e-26 -7.17e-02 2.895496116e+00 2.894787867e+00 5.3e-25 0.00

42 6.4e-24 4.7e-25 1.5e-26 3.56e-01 2.896958288e+00 2.896575882e+00 1.3e-25 0.00

43 1.1e-23 1.6e-25 6.6e-27 -4.71e-02 2.899796396e+00 2.899186023e+00 4.2e-26 0.00

44 2.6e-24 4.2e-26 1.1e-27 3.80e-01 2.901144025e+00 2.900826139e+00 1.0e-26 0.00

45 1.2e-24 1.5e-26 4.8e-28 -3.93e-02 2.903739795e+00 2.903228499e+00 3.2e-27 0.00

46 3.0e-25 8.5e-28 4.5e-29 2.39e-01 2.905594920e+00 2.905240326e+00 6.6e-28 0.00

47 8.9e-26 4.0e-28 6.0e-30 -3.37e-03 2.907754518e+00 2.907232163e+00 2.3e-28 0.00

Optimizer terminated. Time: 0.00

the solution found is

[2.3028378190668097e-14, 2.402867139391792e-14, 7.127137945447063e-13]

which violates the sum(p) == 1 constraint.

Is this a mosek bug? i'm using mosek 9.3.18, mosek.jl 1.2.1 adn, mosek tools 0.12.0 on julia 1.6

— Reply to this email directly, view it on GitHub https://github.com/MOSEK/Mosek.jl/issues/212, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACOAE4K722BYO6R3YVX5W63VAY2F3ANCNFSM5REPCP4A . You are receiving this because you are subscribed to this thread.Message ID: @.***>

SobhanMP commented 2 years ago

termination_status(model) returns MOI.OPTIMAL.

ulfworsoe commented 2 years ago

I can see that there is a bug in the termination status reporting in MosekTools.

This should not take too long to fix.

ulfworsoe commented 2 years ago

I have made a fix (github.com/ulfworsoe/MosekTools.jl#. Pull request in progress here: https://github.com/jump-dev/MosekTools.jl/pull/93

SobhanMP commented 2 years ago

thanks!