You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
RELOG/src/model/jumpext.jl

121 lines
3.7 KiB

# This file extends some JuMP functions so that decision variables can be safely
# replaced by (constant) floating point numbers.
using Printf
using JuMP
import JuMP: value, fix, set_name
function value(x::Float64)
return x
end
function fix(x::Float64, v::Float64; force)
return abs(x - v) < 1e-6 || error("Value mismatch: $x != $v")
end
function set_name(::Number, ::String)
# nop
end
function _init(model::JuMP.Model, key::Symbol)::OrderedDict
if !(key in keys(object_dictionary(model)))
model[key] = OrderedDict()
end
return model[key]
end
function _set_names!(model::JuMP.Model)
@info "Setting variable and constraint names..."
time_varnames = @elapsed begin
_set_names!(object_dictionary(model))
end
@info @sprintf("Set names in %.2f seconds", time_varnames)
end
function _set_names!(dict::Dict)
for name in keys(dict)
dict[name] isa AbstractDict || continue
for idx in keys(dict[name])
if dict[name][idx] isa AffExpr
continue
end
idx_str = join(map(string, idx), ",")
set_name(dict[name][idx], "$name[$idx_str]")
end
end
end
"""
_add_pwl_constraints(model, xvar, yvars, xpts, ypts)
Add piecewise-linear constraints to a JuMP model for multiple y variables.
Creates constraints y_i = f_i(x) where each f_i is a piecewise-linear function
defined by the breakpoints (xpts, ypts[:, i]).
# Arguments
- `model`: JuMP model
- `xvar`: The x variable (JuMP variable)
- `yvars`: Vector of y variables (JuMP variables)
- `xpts`: Vector of x values for breakpoints (must be in non-decreasing order)
- `ypts`: Matrix of y values where ypts[i, j] is the y value for the j-th variable
at the i-th breakpoint
# Example
```julia
@variable(model, y1)
@variable(model, y2)
ypts_matrix = [1.5 2.0; 0.0 1.5; 3.0 0.5] # 3 breakpoints, 2 y variables
_add_pwl_constraints(model, x, [y1, y2], [0.0, 1.0, 2.0], ypts_matrix, name="multiPWL")
```
"""
function _add_pwl_constraints(model, xvar, yvars, xpts, ypts)
# Input validation
ypts isa AbstractMatrix || throw(ArgumentError("ypts must be a matrix"))
length(xpts) == size(ypts, 1) ||
throw(ArgumentError("xpts length must match number of rows in ypts"))
length(yvars) == size(ypts, 2) ||
throw(ArgumentError("Number of y variables must match number of columns in ypts"))
length(xpts) >= 1 || throw(ArgumentError("At least one breakpoint is required"))
# Check that xpts is increasing
for i = 2:length(xpts)
xpts[i] > xpts[i-1] || throw(ArgumentError("xpts must be in increasing order"))
end
n_points = length(xpts)
n_yvars = length(yvars)
if n_points == 1
# Single point case: y_j = ypts[1,j], x = xpts[1]
@constraint(model, xvar == xpts[1])
for j = 1:n_yvars
@constraint(model, yvars[j] == ypts[1, j])
end
elseif n_points == 2
# Two points case: single linear segment for each y variable
x1, x2 = xpts[1], xpts[2]
# Linear relationship for each y variable: y_j = y1_j + slope_j * (x-x1)
for j = 1:n_yvars
y1, y2 = ypts[1, j], ypts[2, j]
slope = (y2 - y1) / (x2 - x1)
@constraint(model, yvars[j] == y1 + slope * (xvar - x1))
end
else
# Multiple segments case (3+ points): use SOS2 formulation
λ = @variable(model, [1:n_points], lower_bound = 0, upper_bound = 1)
@constraint(model, λ in SOS2())
@constraint(model, sum(λ) == 1)
@constraint(model, xvar == sum(xpts[i] * λ[i] for i = 1:n_points))
for j = 1:n_yvars
@constraint(model, yvars[j] == sum(ypts[i, j] * λ[i] for i = 1:n_points))
end
end
return
end