Skip to content

Commit

Permalink
WIP: Support MOI.ScalarNonlinearFunction
Browse files Browse the repository at this point in the history
  • Loading branch information
@odow
odow committed Apr 2, 2024
1 parent 5f45fb2 commit 1f6f6c1
Show file tree
Hide file tree
Showing 2 changed files with 199 additions and 10 deletions.
2 changes: 1 addition & 1 deletion Project.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@ MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"

[compat]
Libdl = "<0.0.1, 1.6"
MathOptInterface = "1"
MathOptInterface = "1.26"
Test = "<0.0.1, 1.6"
Xpress_jll = "=8.13.4, =9.3.0"
julia = "1.6"
Expand Down
207 changes: 198 additions & 9 deletions src/MOI/MOI_wrapper.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,13 @@ const CleverDicts = MOI.Utilities.CleverDicts
EQUAL_TO,
)

@enum(ObjectiveType, SINGLE_VARIABLE, SCALAR_AFFINE, SCALAR_QUADRATIC)
@enum(
ObjectiveType,
SINGLE_VARIABLE,
SCALAR_AFFINE,
SCALAR_QUADRATIC,
SCALAR_NONLINEAR,
)

@enum(CallbackState, CB_NONE, CB_GENERIC, CB_LAZY, CB_USER_CUT, CB_HEURISTIC)

Expand Down Expand Up @@ -1092,6 +1098,7 @@ function MOI.supports(
MOI.VariableIndex,
MOI.ScalarAffineFunction{Float64},
MOI.ScalarQuadraticFunction{Float64},
MOI.ScalarNonlinearFunction,
},
}
return true
Expand Down Expand Up @@ -1128,6 +1135,8 @@ function MOI.set(
if model.objective_type == SCALAR_QUADRATIC
# We need to zero out the existing quadratic objective.
@checked Lib.XPRSdelqmatrix(model.inner, -1)
elseif model.objective_type == SCALAR_NONLINEAR
@checked Lib.XPRSnlpdelobjformula(model.inner)
end
num_vars = length(model.variable_info)
# We zero all terms because we want to gurantee that the old terms
Expand Down Expand Up @@ -1184,7 +1193,7 @@ function MOI.set(
::MOI.ObjectiveFunction{F},
f::F,
) where {F<:MOI.ScalarQuadraticFunction{Float64}}
# setting linear part also clears the existing quadratic terms
# setting linear part also clears the existing quadratic and nonlinear terms
MOI.set(
model,
MOI.ObjectiveFunction{MOI.ScalarAffineFunction{Float64}}(),
Expand Down Expand Up @@ -1278,6 +1287,155 @@ function MOI.modify(
return
end

function MOI.set(
model::Optimizer,
::MOI.ObjectiveFunction{MOI.ScalarNonlinearFunction},
f::MOI.ScalarNonlinearFunction,
)
parsed, type, value = Cint(1), Cint[], Cdouble[]
_reverse_polish(model, f, type, value)
reverse!(type)
reverse!(value)
push!(type, Lib.XPRS_TOK_EOF)
push!(value, 0.0)
@show type, value
@checked Lib.XPRSnlpdelobjformula(model.inner)
@checked Lib.XPRSnlpchgobjformula(model.inner, parsed, type, value)
model.objective_type = SCALAR_NONLINEAR
model.is_objective_set = true
return
end

const _FUNCTION_MAP = Dict(
# Handled explicitly: Lib.XPRS_OP_UMINUS
:^ => (Lib.XPRS_TOK_OP, Lib.XPRS_OP_EXPONENT),
:* => (Lib.XPRS_TOK_OP, Lib.XPRS_OP_MULTIPLY),
:/ => (Lib.XPRS_TOK_OP, Lib.XPRS_OP_DIVIDE),
:+ => (Lib.XPRS_TOK_OP, Lib.XPRS_OP_PLUS),
:- => (Lib.XPRS_TOK_OP, Lib.XPRS_OP_MINUS),
# const XPRS_DEL_COMMA = 1
# const XPRS_DEL_COLON = 2
:log => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_LOG),
:log10 => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_LOG10),
# const XPRS_IFUN_LN = 15
:exp => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_EXP),
:abs => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ABS),
:sqrt => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_SQRT),
:sin => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_SIN),
:cos => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_COS),
:tan => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_TAN),
:asin => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ARCSIN),
:acos => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ARCCOS),
:atan => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ARCTAN),
:max => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_MIN),
:min => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_MAX),
# const XPRS_IFUN_PWL = 35
# Handled explicitly: XPRS_IFUN_SUM
# Handled explicitly: XPRS_IFUN_PROD
:sign => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_SIGN),
:erf => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ERF),
:erfc => (Lib.XPRS_TOK_IFUN, Lib.XPRS_IFUN_ERFC),
)

function _reverse_polish(
model::Optimizer,
f::MOI.ScalarNonlinearFunction,
type::Vector{Cint},
value::Vector{Cdouble},
)
if f.head == :- && length(f.args) == 1
# Special handling for unary negation
push!(type, Lib.XPRS_TOK_OP)
push!(value, Lib.XPRS_OP_UMINUS)
for arg in reverse(f.args)
_reverse_polish(model, arg, type, value)
end
return
elseif f.head in (:+, :*) && length(f.args) != 2
# Special handling for non-binary sum and product
push!(type, Lib.XPRS_TOK_IFUN)
push!(value, f.head == :+ ? Lib.XPRS_IFUN_SUM : Lib.XPRS_IFUN_PROD)
push!(type, Lib.XPRS_TOK_LB)
push!(value, 0.0)
for arg in reverse(f.args)
_reverse_polish(model, arg, type, value)
end
push!(type, Lib.XPRS_TOK_RB)
push!(value, 0.0)
return
end
ret = get(_FUNCTION_MAP, f.head, nothing)
if ret === nothing
throw(MOI.UnsupportedNonlinearOperator(f.head))
elseif ret[1] == Lib.XPRS_TOK_OP
push!(type, ret[1])
push!(value, ret[2])
for arg in reverse(f.args)
_reverse_polish(model, arg, type, value)
end
else
@assert ret[1] == Lib.XPRS_TOK_IFUN
push!(type, ret[1])
push!(value, ret[2])
# XPRS_TOK_LB is not needed. Implied by XPRS_TOK_IFUN
for arg in reverse(f.args)
_reverse_polish(model, arg, type, value)
end
push!(type, Lib.XPRS_TOK_RB)
push!(value, 0.0)
end
return
end

function _reverse_polish(
::Optimizer,
f::Real,
type::Vector{Cint},
value::Vector{Cdouble},
)
push!(type, Lib.XPRS_TOK_CON)
push!(value, Cdouble(f))
return
end

function _reverse_polish(
model::Optimizer,
x::MOI.VariableIndex,
type::Vector{Cint},
value::Vector{Cdouble},
)
push!(type, Lib.XPRS_TOK_COL)
push!(value, _info(model, x).column - 1)
return
end

function _reverse_polish(
model::Optimizer,
f::MOI.AbstractScalarFunction,
type::Vector{Cint},
value::Vector{Cdouble},
)
_reverse_polish(model, convert(MOI.ScalarNonlinearFunction, f), type, value)
return
end

function MOI.get(
model::Optimizer,
::MOI.ObjectiveFunction{MOI.ScalarNonlinearFunction},
)
p_ntypes = Ref{Cint}()
@checked(
Lib.XPRSnlpgetobjformula(model.inner, 1, 0, p_ntypes, C_NULL, C_NULL)
)
maxtype = p_ntypes[]
type, value = zeros(Cint, maxtype), zeros(Cdouble, maxtype)
@checked(
Lib.XPRSnlpgetobjformula(model.inner, 1, maxtype, p_ntypes, type, value)
)
error("FIXME")
return
end

##
## VariableIndex-in-Set constraints.
##
Expand Down Expand Up @@ -2890,11 +3048,13 @@ function MOI.optimize!(model::Optimizer)
_set_MIP_start(model)
end
start_time = time()
if is_mip(model)
@checked Lib.XPRSmipoptimize(model.inner, model.solve_method)
else
@checked Lib.XPRSlpoptimize(model.inner, model.solve_method)
end
solvestatus, solstatus = Ref{Cint}(), Ref{Cint}()
@checked Lib.XPRSoptimize(
model.inner,
model.solve_method,
solvestatus,
solstatus,
)
model.cached_solution.solve_time = time() - start_time
check_cb_exception(model)
# Should be almost a no-op if not needed. Might have minor overhead due to
Expand Down Expand Up @@ -3366,8 +3526,9 @@ end
function MOI.get(model::Optimizer, attr::MOI.ObjectiveValue)
_throw_if_optimize_in_progress(model, attr)
MOI.check_result_index_bounds(model, attr)
attr = is_mip(model) ? Lib.XPRS_MIPOBJVAL : Lib.XPRS_LPOBJVAL
return @_invoke Lib.XPRSgetdblattrib(model.inner, attr, _)::Float64
return @_invoke(
Lib.XPRSgetdblattrib(model.inner, Lib.XPRS_OBJVAL, _)::Float64
)
end

#=
Expand Down Expand Up @@ -4611,3 +4772,31 @@ function MOI.set(model::Optimizer, ::MOI.AbsoluteGapTolerance, value::Float64)
setcontrol!(model.inner, "XPRS_MIPABSSTOP", value)
return
end

#=
ScalarNonlinearFunction
=#

# function MOI.supports_constraint(
# ::Optimizer,
# ::MOI.ScalarNonlinearFunction,
# ::Union{
# MOI.LessThan{Float64},
# MOI.GreaterThan{Float64},
# MOI.EqualTo{Float64},
# },
# )
# return true
# end

# function MOI.add_constraint(
# ::Optimizer,
# ::MOI.ScalarNonlinearFunction,
# ::Union{
# MOI.LessThan{Float64},
# MOI.GreaterThan{Float64},
# MOI.EqualTo{Float64},
# },
# )
# return true
# end

0 comments on commit 1f6f6c1

Please sign in to comment.