Add constructor kwarg set_inner_silent

[sandyspiers]

Adds a keyword argument set_inner_silent to model constructor, which can be used to avoid MOA setting the inner optimizer silent.

Use case: I faced this issue when using Gurobi as in the lower level optimizer, where I could not find a way to turn back on the output. It seems like any output parameters are ignored and overwritten when optimize! is called (I guess because of Gurobi's lazy model generation?).

A workaround to get the output is if you call optimize! first, then set OutputFlag=1, then re-optimize. But this is unreasonable for large models. An example is shown below.

using JuMP, Gurobi, MultiObjectiveAlgorithms
MOA = MultiObjectiveAlgorithms

@info "First solve..."
n = 100
opt = optimizer_with_attributes(() -> MOA.Optimizer(Gurobi.Optimizer))
mdl = Model(opt)
@variable(mdl, x[1:n], Bin)
@objective(mdl, Max, [rand(n)' * x, rand(n)' * x])
set_optimizer_attribute(mdl, "OutputFlag", 1)
optimize!(mdl)

println("\n")
@info "Second attempt...."
set_optimizer_attribute(mdl, "OutputFlag", 1)
optimize!(mdl)

println("\n")
@info "With set_inner_silent=false"
n = 100
opt = optimizer_with_attributes(
    () -> MOA.Optimizer(Gurobi.Optimizer; set_inner_silent = false),
)
mdl = Model(opt)
@variable(mdl, x[1:n], Bin)
@objective(mdl, Max, [rand(n)' * x, rand(n)' * x])
set_optimizer_attribute(mdl, "OutputFlag", 1)
optimize!(mdl)

Which has the following output:

[ Info: First solve...
Set parameter WLSAccessID
Set parameter WLSSecret
Set parameter LicenseID to value 2461048
Academic license 2461048 - for non-commercial use only - registered to sa___@curtin.edu.au
Set parameter OutputFlag to value 1
----------------------------------------------
        MultiObjectiveAlgorithms.jl
----------------------------------------------
Algorithm: Nothing
----------------------------------------------
solve #     Obj. 1       Obj. 2       Time
----------------------------------------------
    1    4.67116e+01  5.29008e+01  3.58649e+00
    2    4.67116e+01  5.29008e+01  3.67085e+00
    3    4.67116e+01  5.29008e+01  4.05418e+00
    4    4.67116e+01  5.29008e+01  4.05596e+00
----------------------------------------------
termination_status: OPTIMAL
result_count: 1

Total solve time:          5.05048e+00
Time spent in subproblems: 7.92038e-01 (16%)
Number of subproblems:     6
----------------------------------------------


[ Info: Second attempt....
Set parameter OutputFlag to value 1
----------------------------------------------
        MultiObjectiveAlgorithms.jl
----------------------------------------------
Algorithm: Nothing
----------------------------------------------
solve #     Obj. 1       Obj. 2       Time
----------------------------------------------
Gurobi Optimizer version 13.0.0 build v13.0.0rc1 (linux64 - "Arch Linux")

CPU model: AMD Ryzen 7 7730U with Radeon Graphics, instruction set [SSE2|AVX|AVX2]
Thread count: 8 physical cores, 16 logical processors, using up to 16 threads

Academic license 2461048 - for non-commercial use only - registered to sa___@curtin.edu.au
Optimize a model with 0 rows, 100 columns and 0 nonzeros (Max)
Model fingerprint: 0x59950ce6
Model has 100 linear objective coefficients

*** OUTPUT TRUNCATED ***

Explored 0 nodes (0 simplex iterations) in 0.00 seconds (0.00 work units)
Thread count was 1 (of 16 available processors)

Solution count 1: 52.9008

Optimal solution found (tolerance 1.00e-04)
Best objective 5.290076734093e+01, best bound 5.290076734093e+01, gap 0.0000%

User-callback calls 30, time in user-callback 0.00 sec
Set parameter TimeLimit to value 1e+100
----------------------------------------------
termination_status: OPTIMAL
result_count: 1

Total solve time:          1.90709e-02
Time spent in subproblems: 1.27451e-02 (67%)
Number of subproblems:     6
----------------------------------------------


[ Info: With set_inner_silent=false
Set parameter WLSAccessID
Set parameter WLSSecret
Set parameter LicenseID to value 2461048
Academic license 2461048 - for non-commercial use only - registered to sa___@curtin.edu.au
Set parameter OutputFlag to value 1
Set parameter OutputFlag to value 1
----------------------------------------------
        MultiObjectiveAlgorithms.jl
----------------------------------------------
Algorithm: Nothing
----------------------------------------------
solve #     Obj. 1       Obj. 2       Time
----------------------------------------------
Set parameter OutputFlag to value 1
Gurobi Optimizer version 13.0.0 build v13.0.0rc1 (linux64 - "Arch Linux")

CPU model: AMD Ryzen 7 7730U with Radeon Graphics, instruction set [SSE2|AVX|AVX2]
Thread count: 8 physical cores, 16 logical processors, using up to 16 threads

Academic license 2461048 - for non-commercial use only - registered to sa___@curtin.edu.au
Optimize a model with 0 rows, 100 columns and 0 nonzeros (Max)
Model fingerprint: 0x892d8ed3

*** OUTPUT TRUNCATED ***

Solution count 1: 47.9289

Optimal solution found (tolerance 1.00e-04)
Best objective 4.792890566188e+01, best bound 4.792890566188e+01, gap 0.0000%

User-callback calls 30, time in user-callback 0.00 sec
Set parameter TimeLimit to value 1e+100
----------------------------------------------
termination_status: OPTIMAL
result_count: 1

Total solve time:          1.37079e-02
Time spent in subproblems: 1.00780e-02 (74%)
Number of subproblems:     6
----------------------------------------------

Thanks!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.11%. Comparing base (2f32152) to head (3f3199e).
⚠️ Report is 4 commits behind head on master.

❗ There is a different number of reports uploaded between BASE (2f32152) and HEAD (3f3199e). Click for more details.

HEAD has 3 uploads less than BASE

Flag	BASE (2f32152)	HEAD (3f3199e)
	4	1

Additional details and impacted files

@@             Coverage Diff             @@
##           master     #180       +/-   ##
===========================================
- Coverage   99.92%   88.11%   -11.81%     
===========================================
  Files          12       12               
  Lines        1280     1279        -1     
===========================================
- Hits         1279     1127      -152     
- Misses          1      152      +151     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sandyspiers]

On second thought, I think I would prefer if there was an attribute, so that the following would work:

opt = optimizer_with_attributes(
    () -> MOA.Optimizer(Gurobi.Optimizer),
    MOA.SilentInner() => false,
)

I guess this would require moving

MultiObjectiveAlgorithms.jl/src/MultiObjectiveAlgorithms.jl

Line 186 in 2f32152

if MOI.supports(inner, MOI.Silent())

into the optimizer_inner! function.

Is either option preferred by the maintainers?

[odow]

The argument option please. Otherwise I can do this tomorrow. I had assumed the inner solves are too noisy, but I guess it can hep debugging

[sandyspiers]

Sounds good, I am happy to stick with the argument. Are any other changes required?

[odow]

Oh no, I mistyped. I meant the attribute. Sorry on phone