Sort based on optimization sense (#152)

Author: odow

README.md

@@ -89,6 +89,11 @@ Query the number of scalar subproblems that were solved using
 
  * `MOA.SubproblemCount()`
 
+## Solution ordering
+
+Results are lexicograhically ordered by their objective vectors. The order
+depends on the objective sense. The first result is best.
+
 ## Ideal point
 
 By default, MOA will compute the ideal point, which can be queried using the

src/MultiObjectiveAlgorithms.jl

@@ -42,14 +42,15 @@ function dominates(
     end
 end
 
-_sort!(solutions::Vector{SolutionPoint}) = sort!(solutions; by = x -> x.y)
+function _sort!(solutions::Vector{SolutionPoint}, sense::MOI.OptimizationSense)
+    return sort!(solutions; by = x -> x.y, rev = sense == MOI.MAX_SENSE)
+end
 
 function filter_nondominated(
     sense,
     solutions::Vector{SolutionPoint};
     atol::Float64 = 1e-6,
 )
-    _sort!(solutions)
     nondominated_solutions = SolutionPoint[]
     for candidate in solutions
         if any(test -> dominates(sense, test, candidate; atol), solutions)
@@ -60,6 +61,7 @@ function filter_nondominated(
             push!(nondominated_solutions, candidate)
         end
     end
+    _sort!(nondominated_solutions, sense)
     return nondominated_solutions
 end
 
@@ -681,7 +683,7 @@ function _optimize!(model::Optimizer)
     model.termination_status = status
     if solutions !== nothing
         model.solutions = solutions
-        _sort!(model.solutions)
+        _sort!(model.solutions, MOI.get(model, MOI.ObjectiveSense()))
     end
     if MOI.get(model, ComputeIdealPoint())
         _compute_ideal_point(model, start_time)

test/algorithms/Chalmet.jl

@@ -105,6 +105,7 @@ function test_knapsack_max()
         [0, 1, 1, 1, 1, 0, 1, 0, 1, 1] => [3043, 4627],
         [1, 0, 1, 1, 1, 0, 1, 1, 0, 1] => [3395, 3817],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == length(results)
     for (i, (x_sol, y_sol)) in enumerate(results)
         @test ≈(x_sol, MOI.get(model, MOI.VariablePrimal(i), x); atol = 1e-6)

test/algorithms/Dichotomy.jl

@@ -98,6 +98,8 @@ function test_moi_bolp_1_maximize()
     @test MOI.get(model, MOI.ResultCount()) == 3
     X = [[1.0, 0.25], [0.5, 0.5], [0.0, 1.0]]
     Y = [[-2.25, -1.25], [-1.5, -1.5], [-1.0, -2.5]]
+    reverse!(X)
+    reverse!(Y)
     for i in 1:3
         @test MOI.get(model, MOI.PrimalStatus(i)) == MOI.FEASIBLE_POINT
         @test MOI.get(model, MOI.DualStatus(i)) == MOI.NO_SOLUTION
@@ -227,6 +229,7 @@ function test_biobjective_knapsack()
         [948.0, 939.0] => [1, 2, 3, 5, 6, 8, 10, 11, 15, 16, 17],
         [955.0, 906.0] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
     ]
+    reverse!(results)
     for i in 1:MOI.get(model, MOI.ResultCount())
         x_sol = MOI.get(model, MOI.VariablePrimal(i), x)
         @test results[i][2] == findall(elt -> elt > 0.9, x_sol)

test/algorithms/EpsilonConstraint.jl

@@ -69,6 +69,7 @@ function test_biobjective_knapsack()
         [950, 915] => [1, 2, 5, 6, 8, 9, 10, 11, 15, 16, 17],
         [956, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == 9
     for i in 1:MOI.get(model, MOI.ResultCount())
         x_sol = MOI.get(model, MOI.VariablePrimal(i), x)
@@ -113,6 +114,7 @@ function test_biobjective_knapsack_atol()
         [949, 915] => [1, 2, 5, 6, 8, 9, 10, 11, 15, 16, 17],
         [955, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == 9
     for i in 1:MOI.get(model, MOI.ResultCount())
         x_sol = MOI.get(model, MOI.VariablePrimal(i), x)
@@ -154,6 +156,7 @@ function test_biobjective_knapsack_atol_large()
         [948, 939] => [1, 2, 3, 5, 6, 8, 10, 11, 15, 16, 17],
         [955, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == 4
     for i in 1:MOI.get(model, MOI.ResultCount())
         x_sol = MOI.get(model, MOI.VariablePrimal(i), x)
@@ -238,6 +241,7 @@ function test_biobjective_knapsack_min_solution_limit()
         [943, 940] => [2, 3, 5, 6, 8, 9, 10, 11, 15, 16, 17],
         [955, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == 3
     for i in 1:MOI.get(model, MOI.ResultCount())
         x_sol = MOI.get(model, MOI.VariablePrimal(i), x)

test/algorithms/Lexicographic.jl

@@ -94,6 +94,7 @@ function test_knapsack_default()
         [1, 0, 1] => [1, 0, 0, 1],
         [1, 1, 0] => [1, 1, 0, 0],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == 3
     for i in 1:MOI.get(model, MOI.ResultCount())
         X = round.(Int, MOI.get(model, MOI.VariablePrimal(i), x))
@@ -247,6 +248,7 @@ function test_knapsack_5_objectives()
         [1, 0, 1, 0, 2] => [1, 0, 1, 0],
         [1, 1, 0, 0, 2] => [1, 1, 0, 0],
     ]
+    reverse!(results)
     for i in 1:MOI.get(model, MOI.ResultCount())
         X = round.(Int, MOI.get(model, MOI.VariablePrimal(i), x))
         Y = round.(Int, MOI.get(model, MOI.ObjectiveValue(i)))

test/algorithms/RandomWeighting.jl

@@ -124,6 +124,7 @@ function test_knapsack_max()
         [0, 1, 1, 1, 1, 0, 1, 0, 1, 1] => [3043, 4627],
         [1, 0, 1, 1, 1, 0, 1, 1, 0, 1] => [3395, 3817],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == length(results)
     for (i, (x_sol, y_sol)) in enumerate(results)
         @test ≈(x_sol, MOI.get(model, MOI.VariablePrimal(i), x); atol = 1e-6)

test/problems.jl

@@ -113,6 +113,7 @@ function test_problem_knapsack_max_p3(model)
         [0, 1, 1, 1, 1, 0, 1, 0, 1, 1] => [3042, 4627, 3189],
         [1, 0, 1, 1, 1, 0, 1, 1, 0, 1] => [3394, 3817, 3408],
     ]
+    reverse!(results)
     N = MOI.get(model, MOI.ResultCount())
     @assert N == length(results)
     for i in 1:length(results)
@@ -223,6 +224,7 @@ function test_problem_knapsack_max_p4(model)
         [0, 1, 1, 0, 1, 1, 1, 1, 1, 0] => [3152, 3232, 3596, 3382],
         [1, 1, 1, 0, 1, 1, 1, 0, 0, 0] => [3269, 2320, 3059, 2891],
     ]
+    reverse!(results)
     @test MOI.get(model, MOI.ResultCount()) == length(results)
     for (i, (x_sol, y_sol)) in enumerate(results)
         @test ≈(x_sol, MOI.get(model, MOI.VariablePrimal(i), x); atol = 1e-6)
@@ -368,7 +370,7 @@ function test_problem_assignment_max_p3(model)
     MOI.set(model, MOI.ObjectiveSense(), MOI.MAX_SENSE)
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
     MOI.optimize!(model)
-    results = reverse([
+    results = [
         [0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 1] => [16, 61, 47],
         [0 0 1 0 0 0 1 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1] => [17, 43, 71],
         [0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1] => [18, 47, 67],
@@ -390,7 +392,7 @@ function test_problem_assignment_max_p3(model)
         [0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0] => [43, 51, 31],
         [0 1 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 0 1 0 0] => [45, 33, 34],
         [0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0] => [50, 40, 32],
-    ])
+    ]
     @test MOI.get(model, MOI.ResultCount()) == length(results)
     @test MOI.get(model, MOA.SubproblemCount()) >= length(results)
     for (i, (x_sol, y_sol)) in enumerate(results)

test/test_utilities.jl

@@ -25,7 +25,8 @@ function test_filter_nondominated()
     x = Dict{MOI.VariableIndex,Float64}()
     solutions = [MOA.SolutionPoint(x, [0, 1]), MOA.SolutionPoint(x, [1, 0])]
     @test MOA.filter_nondominated(MOI.MIN_SENSE, solutions) == solutions
-    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) == solutions
+    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) ==
+          reverse(solutions)
     return
 end
 
@@ -34,7 +35,7 @@ function test_filter_nondominated_sort_in_order()
     solutions = [MOA.SolutionPoint(x, [0, 1]), MOA.SolutionPoint(x, [1, 0])]
     r_solutions = reverse(solutions)
     @test MOA.filter_nondominated(MOI.MIN_SENSE, r_solutions) == solutions
-    @test MOA.filter_nondominated(MOI.MAX_SENSE, r_solutions) == solutions
+    @test MOA.filter_nondominated(MOI.MAX_SENSE, r_solutions) == r_solutions
     return
 end
 
@@ -55,7 +56,7 @@ function test_filter_nondominated_weakly_dominated()
         MOA.SolutionPoint(x, [1, 0]),
     ]
     @test MOA.filter_nondominated(MOI.MIN_SENSE, solutions) == solutions[[1, 3]]
-    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) == solutions[[2, 3]]
+    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) == solutions[[3, 2]]
     solutions = [
         MOA.SolutionPoint(x, [0, 1]),
         MOA.SolutionPoint(x, [0.5, 1]),
@@ -67,7 +68,7 @@ function test_filter_nondominated_weakly_dominated()
     @test MOA.filter_nondominated(MOI.MIN_SENSE, solutions) ==
           solutions[[1, 4, 6]]
     @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) ==
-          solutions[[3, 5, 6]]
+          solutions[[6, 5, 3]]
     return
 end
 
@@ -82,7 +83,7 @@ function test_filter_nondominated_knapsack()
     ]
     result = solutions[[1, 3, 4]]
     @test MOA.filter_nondominated(MOI.MIN_SENSE, solutions) == result
-    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) == result
+    @test MOA.filter_nondominated(MOI.MAX_SENSE, solutions) == reverse(result)
     return
 end
 
@@ -115,7 +116,7 @@ function test_filter_epsilon()
     solutions =
         [MOA.SolutionPoint(x, [1, 1 + 9e-5]), MOA.SolutionPoint(x, [2, 1])]
     new_solutions = MOA.filter_nondominated(MOI.MAX_SENSE, copy(solutions))
-    @test new_solutions == solutions
+    @test new_solutions == reverse(solutions)
     solutions =
         [MOA.SolutionPoint(x, [-1, -1 - 1e-6]), MOA.SolutionPoint(x, [-2, -1])]
     new_solutions = MOA.filter_nondominated(MOI.MIN_SENSE, copy(solutions))