Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add_minimalsurface #182

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
105 changes: 105 additions & 0 deletions src/ADNLPProblems/minimalsurface.jl
Original file line number Diff line number Diff line change
@@ -0,0 +1,105 @@
using Plots
using ADNLPModels, NLPModels, NLPModelsIpopt, DataFrames, LinearAlgebra, Distances, SolverCore, PyPlot
tmigot marked this conversation as resolved.
Show resolved Hide resolved

tmigot marked this conversation as resolved.
Show resolved Hide resolved
function minimalsurface(; n::Int = default_nvar, type::Val{T} = Val(Float64), kwargs...) where {T}
tmigot marked this conversation as resolved.
Show resolved Hide resolved

# domain definition
xmin = T(0.)
xmax = T(1.)
ymin = T(0.)
ymax = T(1.)

# Definition of the mesh
nx = 20 # number of points according to the direction x
ny = 20 # number of points according to the direction y


x_mesh = LinRange(xmin, xmax, nx) # coordinates of the mesh points x
y_mesh = LinRange(ymin, ymax, ny) # coordinates of the mesh points y

v_D = zeros(nx,ny) # Surface matrix initialization
for i in 1:nx
for j in 1:ny
v_D[i, j] = T(1 - (2 * x_mesh[i] - 1)^2)
end
end


function Objective(v)
tmigot marked this conversation as resolved.
Show resolved Hide resolved
v_reshape = reshape(v, (nx, ny)) # vector to matrix conversion
hx = T(1/nx) # step on the x axis
hy = T(1/ny) # step on the y axis
S = T(0.) # sum initialization
# Calculation of the gradient according to x
for i in 1:nx
for j in 1:ny
if i == 1
gi = T((v_reshape[i+1, j] - v_reshape[i, j])/hx)
tmigot marked this conversation as resolved.
Show resolved Hide resolved
elseif i == nx
gi = T((v_reshape[i, j] - v_reshape[i-1, j])/hx)
tmigot marked this conversation as resolved.
Show resolved Hide resolved
else
gi = T((v_reshape[i+1, j] - v_reshape[i, j])/(2 * hx))
tmigot marked this conversation as resolved.
Show resolved Hide resolved
end
# Calculation of the gradient according to x
if j == 1
gj = T((v_reshape[i, j+1] - v_reshape[i, j])/hy)
tmigot marked this conversation as resolved.
Show resolved Hide resolved
elseif j == ny
gj = T((v_reshape[i, j] - v_reshape[i, j-1])/hy)
tmigot marked this conversation as resolved.
Show resolved Hide resolved
else
gj = T((v_reshape[i, j+1] - v_reshape[i, j])/(2 * hy))
tmigot marked this conversation as resolved.
Show resolved Hide resolved
end

s = T(hx * hy * (sqrt(1 + (gi^2) +(gj)^2))) # Approximation of the derivative with the method of rectangles
tmigot marked this conversation as resolved.
Show resolved Hide resolved
S = S + s # Update the value of S
end
end
return(S)
end

function constraints(v)
v_reshape = reshape(v, (nx, ny)) # vector to matrix conversion
c = similar(v_reshape, nx*ny + 2*(nx +ny)) # creating a constraint vector
index = 1
v_L = zeros(T, nx,ny) # Creation of an obstacle called v_L
for i in 1:nx
for j in 1:ny
if norm(x_mesh[i]-(1/2)) ≤ 1/4 && norm(y_mesh[j]-(1/2)) ≤ 1/4
v_L[i, j] = T(1.) # Update the value of v_L according to the values ​​of x and y
end
end
end
for i in 1:nx
for j in 1:ny
c[index] = T(v_reshape[i, j] - v_L[i, j]) # Constraint that the surface must be above the obstruction
tmigot marked this conversation as resolved.
Show resolved Hide resolved
index = index + 1
end
end
for j in 1:ny
c[index] = T(v_reshape[1, j]) # Constraint: when x=1 or x=nx, the surface is set to 0
tmigot marked this conversation as resolved.
Show resolved Hide resolved
index = index + 1
c[index] = T(v_reshape[nx, j]) # Constraint: when x=1 or x=nx, the surface is set to 0
tmigot marked this conversation as resolved.
Show resolved Hide resolved
index = index + 1
end
for i in 1:nx
c[index] = T(v_reshape[i, 1] - 1 + (2 * i -1)^2) # Constraint: when y=1 or y=ny, the surface follows the function " 1 + (2 * x -1)^2 "
tmigot marked this conversation as resolved.
Show resolved Hide resolved
index = index + 1
c[index] = T(v_reshape[i, ny] - 1 + (2 * i -1)^2) # Constraint: when y=1 or y=ny, the surface follows the function " 1 + (2 * x -1)^2 "
tmigot marked this conversation as resolved.
Show resolved Hide resolved
index = index + 1

end
return c
end


lcon = zeros(T, nx * ny + 2 * nx + 2 * ny) # Lower bound all equal to 0
ucon = zeros(T, nx * ny + 2 * nx + 2 * ny) # Creation of the upper bound vector
ucon[1 : ny * nx] = Inf * ones(T, nx * ny) # first part equal to infinity
tmigot marked this conversation as resolved.
Show resolved Hide resolved
ucon[nx * ny + 1 : end] = zeros(T, 2 * nx + 2 * ny) # second part part equal to zero

v = vec(v_D) #convert to vector

nlp = ADNLPModel(Objective, v, constraints, lcon, ucon)
return nlp
tmigot marked this conversation as resolved.
Show resolved Hide resolved
end


114 changes: 114 additions & 0 deletions src/ADNLPProblems/minsurfo.jl
Original file line number Diff line number Diff line change
@@ -0,0 +1,114 @@
# Find the surface with minimal area, given boundary conditions,
# and above an obstacle.

# This is problem 17 in the COPS (Version 3) collection of
# E. Dolan and J. More'
# see "Benchmarking Optimization Software with COPS"
# Argonne National Labs Technical Report ANL/MCS-246 (2004)
# classification OBR2-AN-V-V

using Plots
using ADNLPModels, NLPModels, NLPModelsIpopt, DataFrames, LinearAlgebra, Distances, SolverCore, PyPlot
Comment on lines +10 to +11
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
using Plots
using ADNLPModels, NLPModels, NLPModelsIpopt, DataFrames, LinearAlgebra, Distances, SolverCore, PyPlot


function minsurfo(; n::Int = default_nvar, type::Val{T} = Val(Float64), kwargs...) where {T}

# domain definition
xmin = T(0.)
xmax = T(1.)
ymin = T(0.)
ymax = T(1.)

# Definition of the mesh
nx = 20 # number of points according to the direction x
ny = 20 # number of points according to the direction y
Comment on lines +22 to +23
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there a reason to choose 20? I suggest we do: nx = Int(round(sqrt(n))) and ny = nx



x_mesh = LinRange(xmin, xmax, nx) # coordinates of the mesh points x
y_mesh = LinRange(ymin, ymax, ny) # coordinates of the mesh points y

v_D = zeros(nx,ny) # Surface matrix initialization
for i in 1:nx
for j in 1:ny
v_D[i, j] = T(1 - (2 * x_mesh[i] - 1)^2)
end
end


function f(v)
v_reshape = reshape(v, (nx, ny)) # vector to matrix conversion
hx = T(1/nx) # step on the x axis
hy = T(1/ny) # step on the y axis
S = T(0.) # sum initialization
# Calculation of the gradient according to x
for i in 1:nx
for j in 1:ny
if i == 1
gi = (v_reshape[i+1, j] - v_reshape[i, j])/hx
elseif i == nx
gi = (v_reshape[i, j] - v_reshape[i-1, j])/hx
else
gi = (v_reshape[i+1, j] - v_reshape[i, j])/(2 * hx)
end
# Calculation of the gradient according to x
if j == 1
gj = (v_reshape[i, j+1] - v_reshape[i, j])/hy
elseif j == ny
gj = (v_reshape[i, j] - v_reshape[i, j-1])/hy
else
gj = (v_reshape[i, j+1] - v_reshape[i, j])/(2 * hy)
end

s = hx * hy * (sqrt(1 + (gi^2) +(gj)^2)) # Approximation of the derivative with the method of rectangles
S = S + s # Update the value of S
Comment on lines +61 to +62
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
s = hx * hy * (sqrt(1 + (gi^2) +(gj)^2)) # Approximation of the derivative with the method of rectangles
S = S + s # Update the value of S
S = S + hx * hy * (sqrt(1 + (gi^2) +(gj)^2)) # Approximation of the derivative with the method of rectangles

end
end
return(S)
end

function constraints(v)
v_reshape = reshape(v, (nx, ny)) # vector to matrix conversion
c = similar(v_reshape, nx*ny + 2*(nx +ny)) # creating a constraint vector
index = 1
v_L = zeros(T, nx,ny) # Creation of an obstacle called v_L
for i in 1:nx
for j in 1:ny
if norm(x_mesh[i]-(1/2)) ≤ 1/4 && norm(y_mesh[j]-(1/2)) ≤ 1/4
v_L[i, j] = T(1.) # Update the value of v_L according to the values ​​of x and y
end
end
end
for i in 1:nx
for j in 1:ny
c[index] = v_reshape[i, j] - v_L[i, j] # Constraint that the surface must be above the obstruction
index = index + 1
end
end
for j in 1:ny
c[index] = v_reshape[1, j] # Constraint: when x=1 or x=nx, the surface is set to 0
index = index + 1
c[index] = v_reshape[nx, j] # Constraint: when x=1 or x=nx, the surface is set to 0
index = index + 1
end
for i in 1:nx
c[index] = v_reshape[i, 1] - 1 + (2 * i -1)^2 # Constraint: when y=1 or y=ny, the surface follows the function " 1 + (2 * x -1)^2 "
index = index + 1
c[index] = v_reshape[i, ny] - 1 + (2 * i -1)^2 # Constraint: when y=1 or y=ny, the surface follows the function " 1 + (2 * x -1)^2 "
index = index + 1

end
return c
end


lcon = zeros(T, nx * ny + 2 * nx + 2 * ny) # Lower bound all equal to 0
ucon = zeros(T, nx * ny + 2 * nx + 2 * ny) # Creation of the upper bound vector
ucon[1 : ny * nx] = T(Inf) * ones(T, nx * ny) # first part equal to infinity
ucon[nx * ny + 1 : end] = zeros(T, 2 * nx + 2 * ny) # second part part equal to zero

v = vec(v_D) #convert to vector

return ADNLPModel(f, v, constraints, lcon, ucon)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
return ADNLPModel(f, v, constraints, lcon, ucon)
return ADNLPModels.ADNLPModel(f, v, constraints, lcon, ucon)


end


26 changes: 26 additions & 0 deletions src/Meta/minimalsurface.jl
Original file line number Diff line number Diff line change
@@ -0,0 +1,26 @@
minimalsurface_meta = Dict(
tmigot marked this conversation as resolved.
Show resolved Hide resolved
:nvar => 400,
:variable_nvar => false,
:ncon => 480,
:variable_ncon => false,
:minimize => true,
:name => "minimalsurface",
:has_equalities_only => false,
:has_inequalities_only => false,
:has_bounds => false,
:has_fixed_variables => false,
:objtype => :other,
:contype => :general,
:best_known_lower_bound => -Inf,
:best_known_upper_bound => Inf,
:is_feasible => missing,
:defined_everywhere => missing,
:origin => :unknown,

)
get_minimalsurface_nvar(; n::Integer = default_nvar, kwargs...) = 400
get_minimalsurface_ncon(; n::Integer = default_nvar, kwargs...) = 480
get_minimalsurface_nlin(; n::Integer = default_nvar, kwargs...) = 0
get_minimalsurface_nnln(; n::Integer = default_nvar, kwargs...) = 480
get_minimalsurface_nequ(; n::Integer = default_nvar, kwargs...) = 80
get_minimalsurface_nineq(; n::Integer = default_nvar, kwargs...) = 400
25 changes: 25 additions & 0 deletions src/Meta/minsurfo.jl
Original file line number Diff line number Diff line change
@@ -0,0 +1,25 @@
minsurfo_meta = Dict(
:nvar => 400,
:variable_nvar => false,
:ncon => 480,
:variable_ncon => false,
:minimize => true,
:name => "minsurfo",
:has_equalities_only => false,
:has_inequalities_only => false,
:has_bounds => false,
:has_fixed_variables => false,
:objtype => :other,
:contype => :general,
:best_known_lower_bound => -Inf,
:best_known_upper_bound => Inf,
:is_feasible => missing,
:defined_everywhere => missing,
:origin => :unknown,
)
get_minsurfo_nvar(; n::Integer = default_nvar, kwargs...) = 400
get_minsurfo_ncon(; n::Integer = default_nvar, kwargs...) = 480
get_minsurfo_nlin(; n::Integer = default_nvar, kwargs...) = 0
get_minsurfo_nnln(; n::Integer = default_nvar, kwargs...) = 480
get_minsurfo_nequ(; n::Integer = default_nvar, kwargs...) = 80
get_minsurfo_nineq(; n::Integer = default_nvar, kwargs...) = 400