Fix and add tests for constant offset in objective function (#76)

src/algorithms/Chalmet.jl

@@ -27,12 +27,10 @@ function _solve_constrained_model(
     f = MOI.Utilities.scalarize(model.f)
     g = sum(1.0 * fi for fi in f)
     MOI.set(model.inner, MOI.ObjectiveFunction{typeof(g)}(), g)
-    constraints = [
-        MOI.add_constraint(model.inner, f[1], MOI.LessThan(rhs[1] - 1))
-        MOI.add_constraint(model.inner, f[2], MOI.LessThan(rhs[2] - 1))
-    ]
+    sets = MOI.LessThan.(rhs .- 1)
+    c = MOI.Utilities.normalize_and_add_constraint.(model.inner, f, sets)
     MOI.optimize!(model.inner)
-    MOI.delete.(model, constraints)
+    MOI.delete.(model, c)
     status = MOI.get(model.inner, MOI.TerminationStatus())
     if !_is_scalar_status_optimal(status)
         return status, nothing
@@ -74,7 +72,11 @@ function optimize_multiobjective!(algorithm::Chalmet, model::Optimizer)
     end
     _, y1[2] = _compute_point(model, variables, f2)
     MOI.set(model.inner, MOI.ObjectiveFunction{typeof(f1)}(), f1)
-    y1_constraint = MOI.add_constraint(model.inner, f2, MOI.LessThan(y1[2]))
+    y1_constraint = MOI.Utilities.normalize_and_add_constraint(
+        model.inner,
+        f2,
+        MOI.LessThan(y1[2]),
+    )
     MOI.optimize!(model.inner)
     x1, y1[1] = _compute_point(model, variables, f1)
     MOI.delete(model.inner, y1_constraint)
@@ -90,7 +92,11 @@ function optimize_multiobjective!(algorithm::Chalmet, model::Optimizer)
         return MOI.OPTIMAL, [solutions]
     end
     MOI.set(model.inner, MOI.ObjectiveFunction{typeof(f2)}(), f2)
-    y2_constraint = MOI.add_constraint(model.inner, f1, MOI.LessThan(y2[1]))
+    y2_constraint = MOI.Utilities.normalize_and_add_constraint(
+        model.inner,
+        f1,
+        MOI.LessThan(y2[1]),
+    )
     MOI.optimize!(model.inner)
     x2, y2[2] = _compute_point(model, variables, f2)
     MOI.delete(model.inner, y2_constraint)

src/algorithms/EpsilonConstraint.jl

@@ -104,7 +104,14 @@ function optimize_multiobjective!(
     else
         MOI.GreaterThan{Float64}, left
     end
-    ci = MOI.add_constraint(model, f1, SetType(bound))
+    constant = MOI.constant(f1, Float64)
+    ci = MOI.Utilities.normalize_and_add_constraint(
+        model,
+        f1,
+        SetType(bound);
+        allow_modify_function = true,
+    )
+    bound -= constant
     status = MOI.OPTIMAL
     for _ in 1:n_points
         if _time_limit_exceeded(model, start_time)
@@ -121,9 +128,9 @@ function optimize_multiobjective!(
             push!(solutions, SolutionPoint(X, Y))
         end
         if sense == MOI.MIN_SENSE
-            bound = min(Y[1] - ε, bound - ε)
+            bound = min(Y[1] - constant - ε, bound - ε)
         else
-            bound = max(Y[1] + ε, bound + ε)
+            bound = max(Y[1] - constant + ε, bound + ε)
         end
     end
     MOI.delete(model, ci)

src/algorithms/Hierarchical.jl

@@ -117,7 +117,8 @@ function optimize_multiobjective!(algorithm::Hierarchical, model::Optimizer)
             else
                 MOI.GreaterThan(Y[i] - rtol * abs(Y[i]))
             end
-            push!(constraints, MOI.add_constraint(model, fi, set))
+            ci = MOI.Utilities.normalize_and_add_constraint(model, fi, set)
+            push!(constraints, ci)
         end
     end
     X, Y = _compute_point(model, variables, model.f)

src/algorithms/KirlikSayin.jl

@@ -152,7 +152,11 @@ function optimize_multiobjective!(algorithm::KirlikSayin, model::Optimizer)
         )
         for (i, f_i) in enumerate(scalars)
             if i != k
-                ci = MOI.add_constraint(model.inner, f_i, SetType(ε[i] + δ))
+                ci = MOI.Utilities.normalize_and_add_constraint(
+                    model.inner,
+                    f_i,
+                    SetType(ε[i] + δ),
+                )
                 push!(ε_constraints, ci)
             end
         end
@@ -168,8 +172,11 @@ function optimize_multiobjective!(algorithm::KirlikSayin, model::Optimizer)
         sum_f = sum(1.0 * s for s in scalars)
         MOI.set(model.inner, MOI.ObjectiveFunction{typeof(sum_f)}(), sum_f)
         # Constraint to eliminate weak dominance
-        zₖ_constraint =
-            MOI.add_constraint(model.inner, scalars[k], MOI.EqualTo(zₖ))
+        zₖ_constraint = MOI.Utilities.normalize_and_add_constraint(
+            model.inner,
+            scalars[k],
+            MOI.EqualTo(zₖ),
+        )
         MOI.optimize!(model.inner)
         MOI.delete.(model, ε_constraints)
         MOI.delete(model, zₖ_constraint)

src/algorithms/Lexicographic.jl

@@ -121,7 +121,8 @@ function _solve_in_sequence(
         else
             MOI.GreaterThan(Y - rtol * abs(Y))
         end
-        push!(constraints, MOI.add_constraint(model, f, set))
+        ci = MOI.Utilities.normalize_and_add_constraint(model, f, set)
+        push!(constraints, ci)
     end
     for c in constraints
         MOI.delete(model, c)

src/algorithms/TambyVanderpooten.jl

@@ -143,7 +143,7 @@ function optimize_multiobjective!(
         ε_constraints = Any[]
         for (i, f_i) in enumerate(scalars)
             if i != k
-                ci = MOI.add_constraint(
+                ci = MOI.Utilities.normalize_and_add_constraint(
                     model.inner,
                     f_i,
                     MOI.LessThan{Float64}(u[i] - 1),
@@ -171,8 +171,11 @@ function optimize_multiobjective!(
         y_k = MOI.get(model.inner, MOI.ObjectiveValue())
         sum_f = sum(1.0 * s for s in scalars)
         MOI.set(model.inner, MOI.ObjectiveFunction{typeof(sum_f)}(), sum_f)
-        y_k_constraint =
-            MOI.add_constraint(model.inner, scalars[k], MOI.EqualTo(y_k))
+        y_k_constraint = MOI.Utilities.normalize_and_add_constraint(
+            model.inner,
+            scalars[k],
+            MOI.EqualTo(y_k),
+        )
         MOI.optimize!(model.inner)
         if !_is_scalar_status_optimal(model)
             return status, nothing

test/algorithms/Chalmet.jl

@@ -98,7 +98,7 @@ function test_knapsack_max()
             MOI.VectorAffineTerm(i, MOI.ScalarAffineTerm(C[i, j], x[j])) for
             i in 1:2 for j in 1:n
         ],
-        [0.0, 0.0],
+        [1.0, 0.0],
     )
     MOI.set(model, MOI.ObjectiveSense(), MOI.MAX_SENSE)
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
@@ -109,9 +109,9 @@ function test_knapsack_max()
         0 1 1 1 1 0 1 0 1 1
     ]
     Y_N = Float64[
-        2854 4636
-        3394 3817
-        3042 4627
+        2855 4636
+        3395 3817
+        3043 4627
     ]
     N = MOI.get(model, MOI.ResultCount())
     x_sol = hcat([MOI.get(model, MOI.VariablePrimal(i), x) for i in 1:N]...)

test/algorithms/Dichotomy.jl

@@ -47,7 +47,7 @@ function test_moi_bolp_1()
         model,
         """
 variables: x, y
-minobjective: [2 * x + y, x + 3 * y]
+minobjective: [2 * x + y + 1, x + 3 * y]
 c1: x + y >= 1.0
 c2: 0.5 * x + y >= 0.75
 c3: x >= 0.0
@@ -60,15 +60,15 @@ c4: y >= 0.25
     @test MOI.get(model, MOI.TerminationStatus()) == MOI.OPTIMAL
     @test MOI.get(model, MOI.ResultCount()) == 3
     X = [[0.0, 1.0], [0.5, 0.5], [1.0, 0.25]]
-    Y = [[1.0, 3.0], [1.5, 2.0], [2.25, 1.75]]
+    Y = [[2.0, 3.0], [2.5, 2.0], [3.25, 1.75]]
     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
         @test MOI.get(model, MOI.ObjectiveValue(i)) == Y[i]
         @test MOI.get(model, MOI.VariablePrimal(i), x) == X[i][1]
         @test MOI.get(model, MOI.VariablePrimal(i), y) == X[i][2]
     end
-    @test MOI.get(model, MOI.ObjectiveBound()) == [1.0, 1.75]
+    @test MOI.get(model, MOI.ObjectiveBound()) == [2.0, 1.75]
     return
 end
 
@@ -83,7 +83,7 @@ function test_moi_bolp_1_maximize()
         model,
         """
 variables: x, y
-maxobjective: [-2.0 * x + -1.0 * y, -1.0 * x + -3.0 * y]
+maxobjective: [-2.0 * x + -1.0 * y, -1.0 * x + -3.0 * y + 0.5]
 c1: x + y >= 1.0
 c2: 0.5 * x + y >= 0.75
 c3: x >= 0.0
@@ -96,15 +96,15 @@ c4: y >= 0.25
     @test MOI.get(model, MOI.TerminationStatus()) == MOI.OPTIMAL
     @test MOI.get(model, MOI.ResultCount()) == 3
     X = [[0.0, 1.0], [0.5, 0.5], [1.0, 0.25]]
-    Y = [-[1.0, 3.0], -[1.5, 2.0], -[2.25, 1.75]]
+    Y = [-[1.0, 2.5], -[1.5, 1.5], -[2.25, 1.25]]
     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
         @test MOI.get(model, MOI.ObjectiveValue(i)) == Y[i]
         @test MOI.get(model, MOI.VariablePrimal(i), x) == X[i][1]
         @test MOI.get(model, MOI.VariablePrimal(i), y) == X[i][2]
     end
-    @test MOI.get(model, MOI.ObjectiveBound()) == -[1.0, 1.75]
+    @test MOI.get(model, MOI.ObjectiveBound()) == -[1.0, 1.25]
     return
 end
 

test/algorithms/DominguezRios.jl

@@ -54,7 +54,7 @@ function test_knapsack_min_p3()
             MOI.VectorAffineTerm(i, MOI.ScalarAffineTerm(-C[i, j], x[j]))
             for i in 1:p for j in 1:n
         ],
-        fill(0.0, p),
+        ones(p),
     )
     MOI.set(model, MOI.ObjectiveSense(), MOI.MIN_SENSE)
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
@@ -77,6 +77,7 @@ function test_knapsack_min_p3()
         -2997 -3539 -3509
         -2518 -3866 -3191
     ]
+    Y_N .+= 1
     N = MOI.get(model, MOI.ResultCount())
     x_sol = hcat([MOI.get(model, MOI.VariablePrimal(i), x) for i in 1:N]...)
     @test isapprox(sort(x_sol; dims = 1), sort(X_E'; dims = 1); atol = 1e-6)

test/algorithms/EpsilonConstraint.jl

@@ -40,6 +40,7 @@ function test_biobjective_knapsack()
         Float64,
         [sum(1.0 * p[i] * x[i] for i in 1:length(w)) for p in [p1, p2]]...,
     )
+    f.constants[1] = 1.0
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
     MOI.add_constraint(
         model,
@@ -48,15 +49,15 @@ function test_biobjective_knapsack()
     )
     MOI.optimize!(model)
     results = Dict(
-        [955, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
-        [949, 915] => [1, 2, 5, 6, 8, 9, 10, 11, 15, 16, 17],
-        [948, 939] => [1, 2, 3, 5, 6, 8, 10, 11, 15, 16, 17],
-        [943, 940] => [2, 3, 5, 6, 8, 9, 10, 11, 15, 16, 17],
-        [936, 942] => [1, 2, 3, 5, 6, 10, 11, 12, 15, 16, 17],
-        [935, 947] => [2, 5, 6, 8, 9, 10, 11, 12, 15, 16, 17],
-        [934, 971] => [2, 3, 5, 6, 8, 10, 11, 12, 15, 16, 17],
-        [927, 972] => [2, 3, 5, 6, 8, 9, 10, 11, 12, 16, 17],
-        [918, 983] => [2, 3, 4, 5, 6, 8, 10, 11, 12, 16, 17],
+        [956, 906] => [2, 3, 5, 6, 9, 10, 11, 14, 15, 16, 17],
+        [950, 915] => [1, 2, 5, 6, 8, 9, 10, 11, 15, 16, 17],
+        [949, 939] => [1, 2, 3, 5, 6, 8, 10, 11, 15, 16, 17],
+        [944, 940] => [2, 3, 5, 6, 8, 9, 10, 11, 15, 16, 17],
+        [937, 942] => [1, 2, 3, 5, 6, 10, 11, 12, 15, 16, 17],
+        [936, 947] => [2, 5, 6, 8, 9, 10, 11, 12, 15, 16, 17],
+        [935, 971] => [2, 3, 5, 6, 8, 10, 11, 12, 15, 16, 17],
+        [928, 972] => [2, 3, 5, 6, 8, 9, 10, 11, 12, 16, 17],
+        [919, 983] => [2, 3, 4, 5, 6, 8, 10, 11, 12, 16, 17],
     )
     @test MOI.get(model, MOI.ResultCount()) == 9
     for i in 1:MOI.get(model, MOI.ResultCount())

test/algorithms/Hierarchical.jl

@@ -46,11 +46,15 @@ function test_knapsack()
     MOI.add_constraint.(model, x, MOI.LessThan(1.0))
     MOI.set(model, MOI.ObjectiveSense(), MOI.MAX_SENSE)
     f = MOI.Utilities.operate(vcat, Float64, P * x...)
+    f.constants[4] = 1_000.0
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
     MOI.add_constraint(model, sum(1.0 * x[i] for i in 1:4), MOI.LessThan(2.0))
     MOI.optimize!(model)
+    @test MOI.get(model, MOI.ResultCount()) == 1
     x_sol = MOI.get(model, MOI.VariablePrimal(), x)
     @test ≈(x_sol, [0.9, 0, 0.9, 0.2]; atol = 1e-3)
+    y_sol = MOI.get(model, MOI.ObjectiveValue())
+    @test ≈(y_sol, P * x_sol .+ [0.0, 0.0, 0.0, 1_000.0]; atol = 1e-4)
     return
 end
 

test/algorithms/KirlikSayin.jl

@@ -51,7 +51,7 @@ function test_knapsack_min_p3()
             MOI.VectorAffineTerm(i, MOI.ScalarAffineTerm(-C[i, j], x[j]))
             for i in 1:p for j in 1:n
         ],
-        fill(0.0, p),
+        ones(p),
     )
     MOI.set(model, MOI.ObjectiveSense(), MOI.MIN_SENSE)
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
@@ -74,6 +74,7 @@ function test_knapsack_min_p3()
         -2518 -3866 -3191
         -2854 -4636 -3076
     ]
+    Y_N .+= 1
     N = MOI.get(model, MOI.ResultCount())
     x_sol = hcat([MOI.get(model, MOI.VariablePrimal(i), x) for i in 1:N]...)
     @test isapprox(sort(x_sol; dims = 1), sort(X_E'; dims = 1); atol = 1e-6)

test/algorithms/Lexicographic.jl

@@ -36,11 +36,15 @@ function test_knapsack()
     MOI.add_constraint.(model, x, MOI.LessThan(1.0))
     MOI.set(model, MOI.ObjectiveSense(), MOI.MAX_SENSE)
     f = MOI.Utilities.operate(vcat, Float64, P * x...)
+    f.constants[4] = 1_000.0
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
     MOI.add_constraint(model, sum(1.0 * x[i] for i in 1:4), MOI.LessThan(2.0))
     MOI.optimize!(model)
+    @test MOI.get(model, MOI.ResultCount()) == 1
     x_sol = MOI.get(model, MOI.VariablePrimal(), x)
     @test ≈(x_sol, [0.9, 1, 0, 0.1]; atol = 1e-3)
+    y_sol = MOI.get(model, MOI.ObjectiveValue())
+    @test ≈(y_sol, P * x_sol .+ [0.0, 0.0, 0.0, 1_000.0]; atol = 1e-4)
     return
 end
 

test/algorithms/TambyVanderpooten.jl

@@ -51,7 +51,7 @@ function test_knapsack_min_p3()
             MOI.VectorAffineTerm(i, MOI.ScalarAffineTerm(-C[i, j], x[j]))
             for i in 1:p for j in 1:n
         ],
-        fill(0.0, p),
+        ones(p),
     )
     MOI.set(model, MOI.ObjectiveSense(), MOI.MIN_SENSE)
     MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
@@ -74,6 +74,7 @@ function test_knapsack_min_p3()
         -2518 -3866 -3191
         -2854 -4636 -3076
     ]
+    Y_N .+= 1
     N = MOI.get(model, MOI.ResultCount())
     x_sol = hcat([MOI.get(model, MOI.VariablePrimal(i), x) for i in 1:N]...)'
     y_sol = hcat([MOI.get(model, MOI.ObjectiveValue(i)) for i in 1:N]...)'