Update DominguezRios to delete boxes that are already searched.

[kofgokhan]

No description provided.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (209ff1c) to head (2ea7d5f).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #166   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           12        12           
  Lines         1206      1213    +7     
=========================================
+ Hits          1206      1213    +7     

☔ 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.

[odow]

I'd write it like this:

diff --git a/src/algorithms/DominguezRios.jl b/src/algorithms/DominguezRios.jl
index e5d8d4c..54ca9ae 100644
--- a/src/algorithms/DominguezRios.jl
+++ b/src/algorithms/DominguezRios.jl
@@ -148,6 +148,10 @@ function _update!(
     return
 end
 
+_isapprox(::Nothing, ::_DominguezRiosBox) = false
+
+_isapprox(A::_DominguezRiosBox, B::_DominguezRiosBox) = A.l ≈ B.l && A.u ≈ B.u
+
 function minimize_multiobjective!(algorithm::DominguezRios, model::Optimizer)
     @assert MOI.get(model.inner, MOI.ObjectiveSense()) == MOI.MIN_SENSE
     n = MOI.output_dimension(model.f)
@@ -190,6 +194,7 @@ function minimize_multiobjective!(algorithm::DominguezRios, model::Optimizer)
     solutions = SolutionPoint[]
     k = 0
     status = MOI.OPTIMAL
+    B_prev = Vector{Union{Nothing,_DominguezRiosBox}}(nothing, n)
     while any(!isempty(l) for l in L)
         if (ret = _check_premature_termination(model)) !== nothing
             status = ret
@@ -197,6 +202,11 @@ function minimize_multiobjective!(algorithm::DominguezRios, model::Optimizer)
         end
         i, k = _select_next_box(L, k)
         B = L[k][i]
+        if _isapprox(B_prev[k], B)
+            deleteat!(L[k], i)
+            continue
+        end
+        B_prev[k] = B
         # We're going to scale `w` here by `scale` instead of the usual
         # `1 / max(...)`. It will show up in a few places bbelow.
         w = scale ./ max.(1, B.u - yI)

[odow]

I guess the alternative is to change the _update! function so that we don't keep adding previously searched boxes?