Update JuMP syntax to use the new nonlinear interface (new)

Project.toml

@@ -14,7 +14,7 @@ SpecialFunctions = "276daf66-3868-5448-9aa4-cd146d93841b"
 [compat]
 DataFrames = "1"
 JLD2 = "0.4"
-JuMP = "~0.19, ~0.20, ~0.21, ~0.22, 0.23, 1"
+JuMP = "^1.15"
 Requires = "1"
 SpecialFunctions = "2"
 julia = "~1"

src/ADNLPProblems/allinit.jl

@@ -14,21 +14,27 @@ function allinit(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) wh
   lvar = -T(Inf) * ones(T, 4)
   uvar = T(Inf) * ones(T, 4)
 
-  function c!(cx, x)
-    cx[1] = x[2] - 1
-    cx[2] = x[3]
-    cx[3] = x[4] - 2
-    return cx
-  end
+  #function c!(cx, x)
+  #  cx[3] = x[4]
+  #  cx[1] = x[2]
+  #  cx[2] = x[3]
+  #  return cx
+  #end
+  A = T[
+    0 0 0 1;
+    0 1 0 0;
+    0 0 1 0
+  ]
 
   return ADNLPModels.ADNLPModel!(
     f,
     x0,
     lvar,
     uvar,
-    c!,
-    T[0, -1e+10, 0],
-    T[Inf, 1, 0],
+    findnz(sparse(A))...,
+    (cx, x) -> cx,
+    T[2, 1, -1e+10],
+    T[2, Inf, 1],
     name = "allinit",
     minimize = true;
     kwargs...,

src/ADNLPProblems/allinitc.jl

@@ -15,21 +15,24 @@ function allinitc(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) w
   uvar = T(Inf) * ones(T, 4)
 
   function c!(cx, x)
-    cx[1] = x[2] - 1
-    cx[2] = x[3]
-    cx[3] = x[4] - 2
-    cx[4] = x[1]^2 + x[2]^2 - 1
+    cx[1] = x[1]^2 + x[2]^2
     return cx
   end
+  A = T[
+    0 0 0 1;
+    0 1 0 0;
+    0 0 1 0
+  ]
 
   return ADNLPModels.ADNLPModel!(
     f,
     x0,
     lvar,
     uvar,
+    sparse(A),
     c!,
-    T[0, -1e+10, 0, -Inf],
-    T[Inf, 1, 0, 0],
+    T[2, 1, -1e+10, -Inf],
+    T[2, Inf, 1, 1],
     name = "allinitc",
     minimize = true;
     kwargs...,

src/ADNLPProblems/alsotame.jl

@@ -12,19 +12,22 @@ function alsotame(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) w
   function c!(cx, x)
     x, y = x[1], x[2]
     cx[1] = sin(-x + y - 1)
-    cx[2] = x
-    cx[3] = y
     return cx
   end
+  A = T[
+    1 0 0;
+    0 1 0
+  ]
 
   return ADNLPModels.ADNLPModel!(
     f,
     x0,
     lvar,
     uvar,
+    findnz(sparse(A))...,
     c!,
-    T[0, -2, -1.5],
-    T[0, 2, 1.5],
+    T[-2, -1.5, 0],
+    T[2, 1.5, 0],
     name = "alsotame",
     minimize = true;
     kwargs...,

src/ADNLPProblems/avion2.jl

@@ -179,123 +179,13 @@ function avion2(::Val{:nlp}; n::Int = default_nvar, type::Type{T} = Float64, kwa
     1,
     1,
   ]
+  A = sparse([1, 2, 3, 2, 4, 3, 1, 4, 4, 4, 4, 5, 5, 7, 10, 14, 6, 8, 14, 14, 6, 13, 15, 7, 7, 8, 15, 9, 15, 10, 15, 11, 15, 12, 15, 13, 15, 9, 14, 11, 14, 12, 14], [1, 1, 2, 5, 5, 6, 7, 7, 8, 9, 10, 10, 
+  19, 20, 20, 20, 22, 22, 22, 23, 24, 26, 31, 33, 34, 35, 35, 37, 37, 38, 38, 39, 39, 40, 40, 41, 41, 47, 47, 48, 48, 49, 49], T[-0.13, -0.7, -1.0, 1.0, -2.0, 1.0, 1.0, -2.0, -2.0, -1.0, 1.0, -20.0, 1.0, -1.0, -0.043, 0.5, -2.0, -0.137, -1.0, 1.0, 1.0, -300.0, 1.0, -1.0, 1.0, 1.0, -1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, -35.0, 660.0, -200.0, 95.0, -120.0, 70.0], 15, 49)
   function c!(cx, x)
-    SR,
-    LR,
-    PK,
-    EF,
-    SX,
-    LX,
-    SD,
-    SK,
-    ST,
-    SF,
-    LF,
-    AM,
-    CA,
-    CB,
-    SO,
-    SS,
-    IMPDER,
-    IMPK,
-    IMPFUS,
-    QI,
-    PT,
-    MV,
-    MC,
-    MD,
-    PD,
-    NS,
-    VS,
-    CR,
-    PM,
-    DV,
-    MZ,
-    VN,
-    QV,
-    QF,
-    IMPTRAIN,
-    IMPMOT,
-    IMPNMOT,
-    IMPPET,
-    IMPPIL,
-    IMPCAN,
-    IMPSNA,
-    MS,
-    EL,
-    DE,
-    DS,
-    IMPVOIL,
-    NM,
-    NP,
-    NG = x[1],
-    x[2],
-    x[3],
-    x[4],
-    x[5],
-    x[6],
-    x[7],
-    x[8],
-    x[9],
-    x[10],
-    x[11],
-    x[12],
-    x[13],
-    x[14],
-    x[15],
-    x[16],
-    x[17],
-    x[18],
-    x[19],
-    x[20],
-    x[21],
-    x[22],
-    x[23],
-    x[24],
-    x[25],
-    x[26],
-    x[27],
-    x[28],
-    x[29],
-    x[30],
-    x[31],
-    x[32],
-    x[33],
-    x[34],
-    x[35],
-    x[36],
-    x[37],
-    x[38],
-    x[39],
-    x[40],
-    x[41],
-    x[42],
-    x[43],
-    x[44],
-    x[45],
-    x[46],
-    x[47],
-    x[48],
-    x[49]
-    cx[1] = SD - (13 // 100) * SR
-    cx[2] = SX - (7 // 10) * SR
-    cx[3] = LX - LR
-    cx[4] = SF - ST - 2 * SD - 2 * SX - 2 * SK
-    cx[5] = IMPFUS - 20 * SF
-    cx[6] = MD - 2 * MV
-    cx[7] = QF - QI - QV
-    cx[8] = IMPTRAIN - (137 // 1000) * MV
-    cx[9] = IMPNMOT - 35 * NM
-    cx[10] = IMPPET - (43 // 1000) * QI
-    cx[11] = IMPPIL - 200 * NP
-    cx[12] = IMPCAN - 120 * NG
-    cx[13] = IMPSNA - 300 * NS - 400
-    cx[14] = MC - MV + 95 * NP + 70 * NG + 660 * NM + (1 // 2) * QI - 380
-    cx[15] = MZ - IMPTRAIN + IMPNMOT + IMPPET + IMPPIL + IMPCAN + IMPSNA + 290
     return cx
   end
-  lcon = zeros(T, 15)
-  ucon = zeros(T, 15)
+  lcon = vcat(zeros(T, 12), 400, 380, -290)
+  ucon = vcat(zeros(T, 12), 400, 380, -290)
   lvar = T[
     10,
     0,
@@ -399,7 +289,7 @@ function avion2(::Val{:nlp}; n::Int = default_nvar, type::Type{T} = Float64, kwa
     2,
   ]
 
-  return ADNLPModels.ADNLPModel!(f, x0, lvar, uvar, c!, lcon, ucon, name = "avion2"; kwargs...)
+  return ADNLPModels.ADNLPModel!(f, x0, lvar, uvar, findnz(A)..., c!, lcon, ucon, name = "avion2"; kwargs...)
 end
 
 function avion2(::Val{:nls}; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) where {T}

src/ADNLPProblems/booth.jl

@@ -6,17 +6,20 @@ function booth(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) wher
   end
   x0 = zeros(T, 2)
 
+  A = T[
+    1 2;
+    2 1
+  ]
   function c!(cx, x)
-    cx[1] = x[1] + 2 * x[2] - 7
-    cx[2] = 2 * x[1] + x[2] - 5
     return cx
   end
   return ADNLPModels.ADNLPModel!(
     f,
     x0,
+    findnz(sparse(A))...,
     c!,
-    zeros(T, 2),
-    zeros(T, 2),
+    T[7, 5],
+    T[7, 5],
     minimize = true,
     name = "booth";
     kwargs...,

src/ADNLPProblems/bqp1var.jl

@@ -6,7 +6,6 @@ function bqp1var(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) wh
   end
   x0 = T[0.25,]
   function c!(cx, x)
-    cx[1] = x[1]
     return cx
   end
   lcon = T[0.0,]
@@ -15,6 +14,7 @@ function bqp1var(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) wh
   return ADNLPModels.ADNLPModel!(
     f,
     x0,
+    findnz(sparse(ones(T, 1, 1)))...,
     c!,
     lcon,
     ucon,

src/ADNLPProblems/bt1.jl

@@ -7,10 +7,10 @@ function bt1(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) where
   end
   x0 = T[0.08, 0.06]
   function c!(cx, x)
-    cx[1] = x[1]^2 + x[2]^2 - 1
+    cx[1] = x[1]^2 + x[2]^2
     return cx
   end
-  lcon = zeros(T, 1)
-  ucon = zeros(T, 1)
+  lcon = T[1]
+  ucon = T[1]
   return ADNLPModels.ADNLPModel!(f, x0, c!, lcon, ucon, name = "bt1"; kwargs...)
 end

src/ADNLPProblems/camshape.jl

@@ -11,25 +11,37 @@ function camshape(args...; n::Int = default_nvar, type::Type{T} = Float64, kwarg
     return -Ti(R_v * pi / n) * sum(y[i] for i = 1:n)
   end
   function c!(cx, y::V; n = n, R_max = R_max, R_min = R_min, θ = eltype(y)(θ)) where {V}
-    cx[1] = R_max - y[n]
-    cx[2] = y[1] - R_min
-    for i = 1:(n - 1)
-      cx[2 + i] = y[i + 1] - y[i]
-    end
-    cx[n + 2] = -R_min * y[1] - y[1] * y[2] + 2 * R_min * y[2] * cos(θ)
-    cx[n + 3] = -R_min^2 - R_min * y[1] + 2 * R_min * y[1] * cos(θ)
-    cx[n + 4] = -y[n - 1] * y[n] - y[n] * R_max + 2 * y[n - 1] * R_max * cos(θ)
-    cx[n + 5] = -2 * R_max * y[n] + 2 * y[n]^2 * cos(θ)
+    cx[1] = -R_min * y[1] - y[1] * y[2] + 2 * R_min * y[2] * cos(θ) # quadratic
+    cx[2] = -y[n - 1] * y[n] - y[n] * R_max + 2 * y[n - 1] * R_max * cos(θ) # quadratic
+    cx[3] = -2 * R_max * y[n] + 2 * y[n]^2 * cos(θ) # quadratic
     for i = 2:(n - 1)
-      cx[n + 4 + i] = -y[i - 1] * y[i] - y[i] * y[i + 1] + 2 * y[i - 1] * y[i + 1] * cos(θ)
+      cx[2 + i] = -y[i - 1] * y[i] - y[i] * y[i + 1] + 2 * y[i - 1] * y[i + 1] * cos(θ) # quadratic
     end
     return cx
   end
 
+  lcon = vcat(-T(Inf), T(-α * θ) * ones(T, n + 1), -T(Inf) * ones(T, n + 1))
+  ucon = vcat(T(0), T(α * θ) * ones(T, n + 1), zeros(T, n + 1))
+
+  A = zeros(T, n + 2, n)
+  A[2,n] = -1
+  lcon[2] -= R_max
+  ucon[2] -= R_max
+  A[3,1] = 1
+  lcon[3] += R_min
+  ucon[3] += R_min
+  for i = 1:(n - 1)
+    A[3 + i, i + 1] = 1
+    A[3 + i, i] = -1
+  end
+  # cx[n + 3] = -R_min^2 - R_min * y[1] + 2 * R_min * y[1] * cos(θ)
+  A[1, 1] =  -R_min + 2 * R_min * cos(θ)
+  lcon[1] += R_min^2
+  ucon[1] += R_min^2
+
   lvar = T(R_min) * ones(T, n)
   uvar = T(R_max) * ones(T, n)
-  lcon = vcat(T(-α * θ) * ones(T, n + 1), -T(Inf) * ones(T, n + 2))
-  ucon = vcat(T(α * θ) * ones(T, n + 1), zeros(T, n + 2))
+
   x0 = T((R_min + R_max) / 2) * ones(T, n)
-  return ADNLPModels.ADNLPModel!(f, x0, lvar, uvar, c!, lcon, ucon, name = "camshape", ; kwargs...)
+  return ADNLPModels.ADNLPModel!(f, x0, lvar, uvar, findnz(sparse(A))..., c!, lcon, ucon, name = "camshape", ; kwargs...)
 end

src/ADNLPProblems/catenary.jl

@@ -1,15 +1,13 @@
 export catenary
 
-function catenary(args...; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) where {T}
+function catenary(args...; n::Int = default_nvar, type::Type{T} = Float64, Bl = 1, FRACT = 0.6, kwargs...) where {T}
   (n % 3 == 0) || @warn("catenary: number of variables adjusted to be a multiple of 3")
   n = 3 * max(1, div(n, 3))
   (n < 6) || @warn("catenary: number of variables adjusted to be greater or equal to 6")
   n = max(n, 6)
 
   ## Model Parameters
   N = div(n, 3) - 2
-  Bl = 1
-  FRACT = 0.6
   d = Bl * (N + 1) * FRACT
 
   gamma = 9.81
@@ -26,7 +24,7 @@ function catenary(args...; n::Int = default_nvar, type::Type{T} = Float64, kwarg
       cx[i] =
         (x[1 + 3 * i] - x[-2 + 3 * i])^2 +
         (x[2 + 3 * i] - x[-1 + 3 * i])^2 +
-        (x[3 + 3 * i] - x[3 * i])^2 - Bl^2
+        (x[3 + 3 * i] - x[3 * i])^2
     end
     return cx
   end
@@ -38,8 +36,8 @@ function catenary(args...; n::Int = default_nvar, type::Type{T} = Float64, kwarg
   lvar[n - 2] = T(d)
   uvar[n - 2] = T(d)
 
-  lcon = zeros(T, N + 1)
-  ucon = zeros(T, N + 1)
+  lcon = zeros(T, N + 1) .+ Bl^2
+  ucon = zeros(T, N + 1) .+ Bl^2
   x0 = zeros(T, n)
 
   for i = 0:(N + 1)

src/ADNLPProblems/elec.jl

@@ -15,13 +15,13 @@ function elec(; n::Int = default_nvar, type::Type{T} = Float64, kwargs...) where
   # Define the constraints on these points (sum of the square of the coordinates = 1)
   function c!(cx, x; n = n)
     for k = 1:n
-      cx[k] = x[k]^2 + x[n + k]^2 + x[2n + k]^2 - 1
+      cx[k] = x[k]^2 + x[n + k]^2 + x[2n + k]^2
     end
     return cx
   end
 
   # bounds on the constraints
-  lcon = ucon = zeros(T, n)
+  lcon = ucon = ones(T, n)
 
   # building a feasible x0
   range0 = T[i / n for i = 1:n]