Merge branch 'add-marine' of https://github.com/tmigot/PDEOptimizatio…

…nProblems into add-marine

src/marine.jl

@@ -2,12 +2,28 @@ export marine
 
 # Marine Population Dynamics COPS Problem v.0.3.1
 # https://www.mcs.anl.gov/~more//cops/cops3.pdf
+<<<<<<< HEAD
 function marine(args...; n = 400, T = 10.0, kwargs...)
   model = CartesianDiscreteModel((0, T), n)
 
   valuetype = Float64
   reffe = ReferenceFE(lagrangian, valuetype, 1)
   Vfree = TestFESpace(model, reffe; conformity = :H1)
+=======
+# n=400, 800, 1600
+function marine(args...; n = 400, kwargs...)
+  T = 10.0
+  model = CartesianDiscreteModel((0, T), n)
+
+  labels = get_face_labeling(model)
+  add_tag_from_tags!(labels, "diri0", [1]) #initial time condition
+
+  valuetype = Float64
+  reffe = ReferenceFE(lagrangian, valuetype, 1)
+  Vfree = TestFESpace(model, reffe; conformity = :H1)
+  VI = TestFESpace(model, reffe; conformity = :H1, labels = labels, dirichlet_tags = ["diri0"])
+  UI = TrialFESpace(VI, 0.0)
+>>>>>>> 795a1d06d1b16edeafe3d23aef2a206e8ce94b7a
   Ufree = TrialFESpace(Vfree)
   Xpde = MultiFieldFESpace([Vfree, Vfree, Vfree, Vfree, Vfree, Vfree, Vfree, Vfree])
   Ypde = MultiFieldFESpace([Ufree, Ufree, Ufree, Ufree, Ufree, Ufree, Ufree, Ufree])
@@ -21,6 +37,7 @@ function marine(args...; n = 400, T = 10.0, kwargs...)
   conv(u, ∇u) = (∇u ⋅ one(∇u)) ⊙ u
   c(u, v) = conv ∘ (v, ∇(u))
   function res(y, u, p)
+<<<<<<< HEAD
     # y1, y2, y3, y4, y5, y6, y7, y8 = y
     # m1, m2, m3, m4, m5, m6, m7, m8, g1, g2, g3, g4, g5, g6, g7 = u
     # p1, p2, p3, p4, p5, p6, p7, p8 = v
@@ -29,6 +46,29 @@ function marine(args...; n = 400, T = 10.0, kwargs...)
       c(y[j], p[j]) - u[8 + j -1] * y[j - 1] + (u[j] + u[8 + j]) * y[j]
     )dΩ for j=2:7) +
     ∫( c(y[8], p[8]) - u[15] * y[7] + u[8] * y[8] )dΩ
+=======
+    y1, y2, y3, y4, y5, y6, y7, y8 = y
+    m1, m2, m3, m4, m5, m6, m7, m8, g1, g2, g3, g4, g5, g6, g7 = u
+    p1, p2, p3, p4, p5, p6, p7, p8 = p
+    return ∫( c(y1, p1) + p1 * (m1 + g1) * y1 + 
+      c(y2, p2) - p2 * g1 * y1 + p2 * (m2 + g2) * y2 +
+      c(y3, p3) - p3 * g2 * y2 + p3 * (m3 + g3) * y3 +
+      c(y4, p4) - p4 * g3 * y3 + p4 * (m4 + g4) * y4 +
+      c(y5, p5) - p5 * g4 * y4 + p5 * (m5 + g5) * y5 +
+      c(y6, p6) - p6 * g5 * y5 + p6 * (m6 + g6) * y6 +
+      c(y7, p7) - p7 * g6 * y6 + p7 * (m7 + g7) * y7 +
+      c(y8, p8) - p8 * g7 * y7 + p8 * m8 * y8 )dΩ
+    #=
+    return ∫( c(y[1], p[1]) + (u[1] + u[8 + 1]) * y[1] + 
+      c(y[2], p[2]) - u[8 + 2 -1] * y[2 - 1] + (u[2] + u[8 + 2]) * y[2] +
+      c(y[3], p[3]) - u[8 + 3 -1] * y[3 - 1] + (u[3] + u[8 + 3]) * y[3] +
+      c(y[4], p[4]) - u[8 + 4 -1] * y[4 - 1] + (u[4] + u[8 + 4]) * y[4] +
+      c(y[5], p[5]) - u[8 + 5 -1] * y[5 - 1] + (u[5] + u[8 + 5]) * y[5] +
+      c(y[6], p[6]) - u[8 + 6 -1] * y[6 - 1] + (u[6] + u[8 + 6]) * y[6] +
+      c(y[7], p[7]) - u[8 + 7 -1] * y[7 - 1] + (u[7] + u[8 + 7]) * y[7] +
+      c(y[8], p[8]) - u[15] * y[7] + u[8] * y[8] )dΩ
+    =#
+>>>>>>> 795a1d06d1b16edeafe3d23aef2a206e8ce94b7a
   end
 
   trian = Triangulation(model)
@@ -83,16 +123,16 @@ function marine(args...; n = 400, T = 10.0, kwargs...)
     1.0    12.0   198.0   707.0  2562.0  3163.0  3232.0  5566.0
   ]
 
-  objterm = PDEOptimizationProblems.InterpolatedEnergyFETerm(8, 21, zmes, 1, τ, dΩ)
+  objterm = PDEOptimizationProblems.InterpolatedEnergyFETerm(8, 21, zmes, 1, τ, dΩ, 1 / n)
   f = (y, u) -> PDEOptimizationProblems.interpolated_measurement(objterm, y)
 
   ndofs = Gridap.FESpaces.num_free_dofs(Ypde) + Gridap.FESpaces.num_free_dofs(Ycon)
   xin = zeros(ndofs)
   return GridapPDENLPModel(xin, f, trian, Ypde, Ycon, Xpde, Xcon, op, name = "Marine Population Dynamics n=$n")
 end
 
-apinene_meta = Dict(
-  :name => "apinene",
+marine_meta = Dict(
+  :name => "marine",
   :domaindim => UInt8(1),
   :pbtype => :yu,
   :nθ => 0,

test/runtests.jl

@@ -4,7 +4,7 @@ using PDEOptimizationProblems
 include("utils.jl")
 
 # Test that every problem can be instantiated.
-for prob in setdiff(names(PDEOptimizationProblems), [:PDEOptimizationProblems])
+for prob in [:marine] #setdiff(names(PDEOptimizationProblems), [:PDEOptimizationProblems])
   @time begin
     print(prob)
     prob_fn = PDEOptimizationProblems.eval(prob)