Skip to content

Fix #4308: Extract and enforce variable bounds in NonlinearProblem and NonlinearLeastSquaresProblem #4463

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.

Sign up for GitHub

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

Open
utkuyilmaz1903 wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Fix #4308: Extract and enforce variable bounds in NonlinearProblem and NonlinearLeastSquaresProblem #4463

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
65 changes: 59 additions & 6 deletions lib/ModelingToolkitBase/src/problems/nonlinearproblem.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,6 +61,33 @@
return maybe_codegen_scimlfn(expression, NonlinearFunction{iip, spec}, args; kwargs...)
end

function _get_nonlinear_bounds_arrays(unknowns_list, u0)
if !any(hasbounds, unknowns_list)
return nothing, nothing
end

T = u0 === nothing ? Float64 : eltype(u0)
if !(T <: Number)
T = Float64
end

N = length(unknowns_list)
lb = Vector{T}(undef, N)
ub = Vector{T}(undef, N)

for (i, v) in enumerate(unknowns_list)
b = getbounds(v)
if b !== nothing && length(b) >= 2
lb[i] = b[1] !== nothing ? T(b[1]) : typemin(T)
ub[i] = b[2] !== nothing ? T(b[2]) : typemax(T)
else
lb[i] = typemin(T)
ub[i] = typemax(T)
end
end
return lb, ub
end

@fallback_iip_specialize function SciMLBase.NonlinearProblem{iip, spec}(
sys::System, op; expression = Val{false},
check_length = true, check_compatibility = true, kwargs...
Expand All @@ -80,9 +107,24 @@ end

kwargs = process_kwargs(sys; kwargs...)
ptype = getmetadata(sys, ProblemTypeCtx, StandardNonlinearProblem())
args = (; f, u0, p, ptype)

return maybe_codegen_scimlproblem(expression, NonlinearProblem{_iip}, args; kwargs...)
if !haskey(kwargs, :lb) && !haskey(kwargs, :ub)
lb, ub = _get_nonlinear_bounds_arrays(unknowns(sys), u0)
if lb !== nothing && ub !== nothing
kwargs = merge(kwargs, (lb = lb, ub = ub))
end
end

if expression == Val{false}
prob = SciMLBase.NonlinearProblem(f, u0, p; kwargs...)
if !(ptype isa StandardNonlinearProblem)
prob = SciMLBase.remake(prob; problem_type = ptype)
end
return prob
else
args = ptype isa StandardNonlinearProblem ? (; f, u0, p) : (; f, u0, p, ptype)
return maybe_codegen_scimlproblem(expression, NonlinearProblem{_iip}, args; kwargs...)
end
end

@fallback_iip_specialize function SciMLBase.NonlinearLeastSquaresProblem{iip, spec}(
Expand All @@ -100,11 +142,22 @@ end
)

kwargs = process_kwargs(sys; kwargs...)
args = (; f, u0, p)

return maybe_codegen_scimlproblem(
expression, NonlinearLeastSquaresProblem{_iip}, args; kwargs...
)
if !haskey(kwargs, :lb) && !haskey(kwargs, :ub)
lb, ub = _get_nonlinear_bounds_arrays(unknowns(sys), u0)
if lb !== nothing && ub !== nothing
kwargs = merge(kwargs, (lb = lb, ub = ub))
end
end

if expression == Val{false}
return SciMLBase.NonlinearLeastSquaresProblem(f, u0, p; kwargs...)
else
args = (; f, u0, p)
return maybe_codegen_scimlproblem(
expression, NonlinearLeastSquaresProblem{_iip}, args; kwargs...
)
end
end

function check_compatible_system(
Expand Down
39 changes: 39 additions & 0 deletions lib/ModelingToolkitBase/test/nonlinearsystem.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -515,3 +515,42 @@ end
prob = NonlinearProblem(sys, [x => 1.0])
@test prob.problem_type == "A"
end

@testset "Bounds extraction for Nonlinear Problems" begin
@variables x [bounds=(0.0, 1.0)]
@variables y [bounds=(-1.0, 5.0)]
@variables z

eqs = [
0 ~ x^2 - 0.5,
0 ~ y - x,
0 ~ z - y
]

@named sys = System(eqs, [x, y, z], [])
sys = complete(sys)

u0 = [x => 0.5, y => 0.5, z => 0.5]

prob_nl = NonlinearProblem(sys, u0)
prob_ls = NonlinearLeastSquaresProblem(sys, u0)

for p in (prob_nl, prob_ls)
@test p.lb !== nothing
@test p.ub !== nothing

@test eltype(p.lb) === Float64
@test eltype(p.ub) === Float64

unks = unknowns(sys)
expected_lb = map(unks) do v
isequal(v, x) ? 0.0 : (isequal(v, y) ? -1.0 : typemin(Float64))
end
expected_ub = map(unks) do v
isequal(v, x) ? 1.0 : (isequal(v, y) ? 5.0 : typemax(Float64))
end

@test p.lb == expected_lb
@test p.ub == expected_ub
end
end
Loading