Define common problem combos

Author: dpo

Project.toml

@@ -27,7 +27,8 @@ DifferentialEquations = "0c46a032-eb83-5123-abaf-570d42b7fbaa"
 MLDatasets = "eb30cadb-4394-5ae3-aed4-317e484a6458"
 ProximalOperators = "a725b495-10eb-56fe-b38b-717eba820537"
 QuadraticModels = "f468eda6-eac5-11e8-05a5-ff9e497bcd19"
+ShiftedProximalOperators = "d4fd37fa-580c-4e43-9b30-361c21aae263"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["ADNLPModels", "DifferentialEquations", "MLDatasets", "ProximalOperators", "QuadraticModels", "Test"]
+test = ["ADNLPModels", "DifferentialEquations", "MLDatasets", "ProximalOperators", "QuadraticModels", "ShiftedProximalOperators", "Test"]

src/RegularizedProblems.jl

@@ -18,15 +18,26 @@ function __init__()
     include("testset_bpdn.jl")
     include("testset_lrcomp.jl")
     include("testset_matrand.jl")
+    include("testset_nnmf.jl")
+  end
+  @require ShiftedProximalOperators = "d4fd37fa-580c-4e43-9b30-361c21aae263" begin
     include("testset_group_lasso.jl")
   end
   @require ADNLPModels = "54578032-b7ea-4c30-94aa-7cbd1cce6c9a" begin
     @require DifferentialEquations = "0c46a032-eb83-5123-abaf-570d42b7fbaa" begin
       include("fh_model.jl")
+      @require ProximalOperators = "a725b495-10eb-56fe-b38b-717eba820537" begin
+        include("testset_fh.jl")
+      end
     end
   end
   @require MLDatasets = "eb30cadb-4394-5ae3-aed4-317e484a6458" begin
     include("nonlin_svm_model.jl")
+    @require ProximalOperators = "a725b495-10eb-56fe-b38b-717eba820537" begin
+      @require ShiftedProximalOperators = "d4fd37fa-580c-4e43-9b30-361c21aae263" begin
+        include("testset_svm.jl")
+      end
+    end
   end
   @require QuadraticModels = "f468eda6-eac5-11e8-05a5-ff9e497bcd19" begin
     include("qp_rand_model.jl")

src/bpdn_model.jl

@@ -21,8 +21,8 @@ bpdn_data(compound::Int = 1, args...; bounds::Bool = false) =
   bpdn_data(200 * compound, 512 * compound, 10 * compound, args...; bounds = bounds)
 
 """
-    model, nls_model, sol = bpdn_model(args...)
-    model, nls_model, sol = bpdn_model(compound = 1, args...)
+    model, nls_model, sol = bpdn_model(args...; kwargs...)
+    model, nls_model, sol = bpdn_model(compound = 1, args...; kwargs...)
 
 Return an instance of an `NLPModel` and an instance of an `NLSModel` representing
 the same basis-pursuit denoise problem, i.e., the under-determined linear

src/qp_rand_model.jl

@@ -2,17 +2,17 @@ export qp_rand_model
 using .QuadraticModels
 
 """
-    model = qp_rand_model(n; dens = 1.0e-4, convex = false)
+    model = qp_rand_model(n = 100_000; dens = 1.0e-4, convex = false)
 
 Return an instance of a `QuadraticModel` representing
 
    min cᵀx + ½ xᵀHx   s.t.  l ≤ x ≤ u,
 
-with H = A + A' or H = A * A' (see the `convex` keyword argument) where A is a random square matrix with density `dens`, `l = -e - tₗ` and `u = e + tᵤ` where `e` is the vector of ones, and `tₗ` and `tᵤ` are sampled from a uniform distribution between 0 and 1.    
+with H = A + A' or H = A * A' (see the `convex` keyword argument) where A is a random square matrix with density `dens`, `l = -e - tₗ` and `u = e + tᵤ` where `e` is the vector of ones, and `tₗ` and `tᵤ` are sampled from a uniform distribution between 0 and 1.
 
 ## Arguments
 
-* `n :: Int`: size of the problem,
+* `n :: Int`: size of the problem (default: `100_000`).
 
 ## Keyword arguments
 
@@ -23,7 +23,7 @@ with H = A + A' or H = A * A' (see the `convex` keyword argument) where A is a r
 
 An instance of a `QuadraticModel`.
 """
-function qp_rand_model(n::Int; dens::R = 1.0e-4, convex::Bool = false) where {R <: Real}
+function qp_rand_model(n::Int = 100_000; dens::R = 1.0e-4, convex::Bool = false) where {R <: Real}
   @assert 0 < dens ≤ 1
   A = sprandn(R, n, n, dens)
   H = convex ? (A * A') : (A + A')

src/testset_bpdn.jl

@@ -2,21 +2,21 @@
 export setup_bpdn_l0, setup_bpdn_l1, setup_bpdn_B0
 
 function setup_bpdn_l0(args...; kwargs...)
-  model, nls_model, _ = bpdn_model(args...)
+  model, nls_model, _ = bpdn_model(args...; kwargs...)
   λ = norm(grad(model, zeros(model.meta.nvar)), Inf) / 10
-  h = NormL0(λ)
+  h = ProximalOperators.NormL0(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_bpdn_l1(args...; kwargs...)
-  model, nls_model, _ = bpdn_model(args...)
+  model, nls_model, _ = bpdn_model(args...; kwargs...)
   λ = norm(grad(model, zeros(model.meta.nvar)), Inf) / 10
-  h = NormL1(λ)
+  h = ProximalOperators.NormL1(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_bpdn_B0(compound = 1, args...; kwargs...)
-  model, nls_model, _ = bpdn_model(compound, args...)
-  h = IndBallL0(10 * compound)
+  model, nls_model, _ = bpdn_model(compound, args...; kwargs...)
+  h = ProximalOperators.IndBallL0(10 * compound)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end

src/testset_fh.jl

@@ -0,0 +1,14 @@
+# Predefine a set of common problem instances.
+export setup_fh_l0, setup_fh_l1
+
+function setup_fh_l0(; kwargs...)
+  model, nls_model, _ = fh_model(; kwargs...)
+  h = ProximalOperators.NormL0(1.0)
+  return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
+end
+
+function setup_fh_l1(; kwargs...)
+  model, nls_model, _ = fh_model(; kwargs...)
+  h = ProximalOperators.NormL1(10.0)
+  return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
+end

src/testset_group_lasso.jl

@@ -5,7 +5,6 @@ function setup_group_lasso_l12(args...; kwargs...)
   model, nls_model, ng, _, idx = group_lasso_model(; kwargs...)
   idx = [idx[i, :] for i = 1:ng]
   λ = 0.2 * ones(ng)
-  h = GroupNormL2(λ, idx)
+  h = ShiftedProximalOperators.GroupNormL2(λ, idx)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
-

src/testset_lrcomp.jl

@@ -2,15 +2,15 @@
 export setup_lrcomp_rank, setup_lrcomp_nuclear
 
 function setup_lrcomp_rank(args...; kwargs...)
-  model, nls_model, _ = lrcomp_model(args...; kwargs...)
+  model, nls_model, _ = lrcomp_model(; kwargs...)
   λ = 0.1
-  h = Rank(λ)
+  h = ProximalOperators.Rank(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_lrcomp_nuclear(args...; kwargs...)
-  model, nls_model, _ = lrcomp_model(args...; kwargs...)
+  model, nls_model, _ = lrcomp_model(; kwargs...)
   λ = 0.1
-  h = NuclearNorm(λ)
+  h = ProximalOperators.NuclearNorm(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end

src/testset_matrand.jl

@@ -5,27 +5,27 @@ export setup_mit_completion_rank, setup_mit_completion_nuclear
 function setup_random_completion_rank(args...; kwargs...)
   model, nls_model, _ = random_matrix_completion_model(; kwargs...)
   λ = 0.1
-  h = Rank(λ)
+  h = ProximalOperators.Rank(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_random_completion_nuclear(args...; kwargs...)
   model, nls_model, _ = random_matrix_completion_model(; kwargs...)
   λ = 0.1
-  h = NuclearNorm(λ)
+  h = ProximalOperators.NuclearNorm(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_mit_completion_rank(args...; kwargs...)
   model, nls_model, _ = MIT_matrix_completion_model()
   λ = 0.1
-  h = Rank(λ)
+  h = ProximalOperators.Rank(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end
 
 function setup_mit_completion_nuclear(args...; kwargs...)
   model, nls_model, _ = MIT_matrix_completion_model()
   λ = 0.1
-  h = NuclearNorm(λ)
+  h = ProximalOperators.NuclearNorm(λ)
   return RegularizedNLPModel(model, h), RegularizedNLSModel(nls_model, h)
 end

src/testset_nnmf.jl

@@ -0,0 +1,16 @@
+# Predefine a set of common problem instances.
+export setup_nnmf_l0, setup_nnmf_l1
+
+function setup_nnmf_l0(args...; kwargs...)
+  model, nls_model, _, selected = nnmf_model(args...)
+  λ = norm(grad(model, rand(model.meta.nvar)), Inf) / 200
+  h = ProximalOperators.NormL0(λ)
+  return RegularizedNLPModel(model, h, selected), RegularizedNLSModel(nls_model, h, selected)
+end
+
+function setup_nnmf_l1(args...; kwargs...)
+  model, nls_model, _, selected = nnmf_model(args...)
+  λ = norm(grad(model, rand(model.meta.nvar)), Inf) / 100_000
+  h = ProximalOperators.NormL1(λ)
+  return RegularizedNLPModel(model, h, selected), RegularizedNLSModel(nls_model, h, selected)
+end