MBM: optional GP-based M sampling on InfiniteGDP models by dnguyen227 · Pull Request #139 · infiniteopt/DisjunctiveProgramming.jl

dnguyen227 · 2026-08-02T03:41:12Z

Implementation of MBM

using AbstractGPs 

optimize!(model, gdp_method = MBM(HiGHS.Optimizer)) 

MBM(HiGHS.Optimizer, M_sampler = :exact)                        # every support, as before
MBM(HiGHS.Optimizer, M_sampler = GPSampler(budget = 0.1))       # solve 10% of supports
MBM(HiGHS.Optimizer, M_sampler = GPSampler(kappa = 4.0))        # more conservative M

Not sure how to work around needing to make a new extension.

codecov · 2026-08-02T03:47:03Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.54%. Comparing base (a4ebe8a) to head (081e722).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #139      +/-   ##
==========================================
+ Coverage   99.51%   99.54%   +0.02%     
==========================================
  Files          17       19       +2     
  Lines        2061     2178     +117     
==========================================
+ Hits         2051     2168     +117     
  Misses         10       10

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

dnguyen227 · 2026-08-24T14:05:57Z

@pulsipher ready for review

pulsipher

This is better, but I think it can be made less complicated.

pulsipher · 2026-08-25T19:21:37Z

+"""
+struct GPSampler
+    # Typed Any so base needs no AbstractGPs dependency
+    kernel::Any


should use parametric type

pulsipher · 2026-08-25T19:21:51Z

    optimizer::O
    default_M::T
-
+    sampler::Any


should use parametric type

pulsipher · 2026-08-25T19:22:33Z

    optimizer::O
    M::Dict{LogicalVariableRef{M}, Any}
    default_M::T
+    sampler::Any


should be parametric type

pulsipher · 2026-08-25T19:25:55Z

+"""
+function sample_M_values(sampler, objectives, sub, method, support_grids)
+    error("Unrecognized `sampler` value `$(repr(sampler))` for MBM " *
+          "on an infinite model. Use `nothing` to solve an M " *


Instead of nothing let's have a type. We should have an abstract type of AbstractMBMSampler with a default of ExhaustiveSampler() instead of nothing.

pulsipher · 2026-08-25T19:31:40Z

+# Bound info for parameter refs in disjunct constraints: parameter
+# functions report their support extrema, finite parameters a point,
+# other parameters (-Inf, Inf). Returns nothing for variables.
+function _parameter_bound_info(
+    vref::InfiniteOpt.GeneralVariableRef
+    )::Union{Nothing, Tuple{Float64, Float64}}
+    _is_parameter(vref) || return nothing
+    dvref = InfiniteOpt.dispatch_variable_ref(vref)
+    if dvref isa InfiniteOpt.ParameterFunctionRef
+        prefs = InfiniteOpt.parameter_list(dvref)
+        length(prefs) == 1 || return (-Inf, Inf)
+        supports = InfiniteOpt.supports(first(prefs))
+        isempty(supports) && return (-Inf, Inf)
+        func = InfiniteOpt.raw_function(dvref)
+        func_values = [func(s) for s in supports]
+        return (minimum(func_values), maximum(func_values))
+    elseif dvref isa InfiniteOpt.FiniteParameterRef
+        value = InfiniteOpt.parameter_value(dvref)
+        return (value, value)
+    end
+    return (-Inf, Inf)
+end
+
+function DP.set_variable_bound_info(
+    vref::InfiniteOpt.GeneralVariableRef, ::DP.BigM)
+    info = _parameter_bound_info(vref)
+    info === nothing || return info
+    lb = JuMP.has_lower_bound(vref) ? JuMP.lower_bound(vref) : -Inf
+    ub = JuMP.has_upper_bound(vref) ? JuMP.upper_bound(vref) : Inf
+    return lb, ub
+end
+
+# Hull and PSplit require finite bounds that include 0
+function _zero_inclusive_bounds(
+    vref::InfiniteOpt.GeneralVariableRef,
+    info::Union{Nothing, Tuple{Float64, Float64}},
+    method_name::String
+    )::Tuple{Float64, Float64}
+    info === nothing || return (min(0, info[1]), max(0, info[2]))
+    if !JuMP.has_lower_bound(vref) || !JuMP.has_upper_bound(vref)
+        error("Variable $vref must have both lower and upper " *
+              "bounds defined when using the $method_name " *
+              "reformulation.")
+    end
+    return (min(0, JuMP.lower_bound(vref)),
+            max(0, JuMP.upper_bound(vref)))
+end
+
+function DP.set_variable_bound_info(
+    vref::InfiniteOpt.GeneralVariableRef, ::DP.Hull)
+    return _zero_inclusive_bounds(vref,
+        _parameter_bound_info(vref), "Hull")
+end
+
+function DP.set_variable_bound_info(
+    vref::InfiniteOpt.GeneralVariableRef,
+    ::Union{DP.PSplit, DP._PSplit})
+    return _zero_inclusive_bounds(vref,
+        _parameter_bound_info(vref), "PSplit")
+end
+


How is this specific to this PR on MBM?

pulsipher · 2026-08-25T19:37:16Z

+# the supports, so it is deferred until sample_M_values needs it.
+function _support_grids(
+    sub::DP.GDPSubmodel, mini_expr::JuMP.AbstractJuMPScalar)
+    reverse_map = Dict(ws[1] => v for (v, ws) in sub.fwd_map)


there are simpler ways to check, like if the model even has any dependent parameters

pulsipher · 2026-08-25T19:39:31Z

+# Transcribe mini_expr, compute the per-support M values with the
+# method's sampler, and aggregate to a scalar if uniform, else to a
+# parameter function on main.
 function DP.raw_M(


why do we call this raw_M?

pulsipher · 2026-08-25T19:44:48Z

+    )
+    M_vals = Array{Float64}(undef, size(objectives))
+    for I in eachindex(objectives)
+        m = DP.raw_M(sub, objectives[I], method)


Does this make raw_M recursive?

pulsipher · 2026-08-25T19:46:42Z

+    objectives::AbstractArray,
+    sub::DP.GDPSubmodel,
+    method::DP._MBM,
+    support_grids::Function


I don't see why this argument is a function or why it needs to be an argument

pulsipher · 2026-08-25T19:50:10Z

+julia> method = MBM(HiGHS.Optimizer, sampler = GPSampler(kappa = 4.0))
+```
+"""
+struct GPSampler


I don't think this makes sense to have in datatypes

dnguyen227 added 3 commits July 29, 2026 12:33

Initial

a930770

Simplify and code coverage

ed08522

Don't compute grids until we need them

fa4a7ca

dnguyen227 added 7 commits August 2, 2026 13:26

detect_uniform_M arg

5a01011

Bound info for parameter functions in disjunct constraints

55e8b3f

New API

6d3909a

AbstractGPsDisjunctiveProgramming

51b3fd9

More keywords

5f90360

GPSampler config object

3721c7b

datatypes and variable name updates

081e722

pulsipher requested changes Aug 25, 2026

View reviewed changes

Navigation Menu

Uh oh!

MBM: optional GP-based M sampling on InfiniteGDP models - #139

MBM: optional GP-based M sampling on InfiniteGDP models #139
dnguyen227 wants to merge 10 commits into
infiniteopt:masterfrom
dnguyen227:mbm_gp

dnguyen227 commented Aug 2, 2026

Uh oh!

codecov Bot commented Aug 2, 2026 •

edited

Loading

Uh oh!

dnguyen227 commented Aug 24, 2026

Uh oh!

pulsipher left a comment

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

pulsipher Aug 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

dnguyen227 commented Aug 2, 2026

Uh oh!

codecov Bot commented Aug 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

dnguyen227 commented Aug 24, 2026

Uh oh!

pulsipher left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codecov Bot commented Aug 2, 2026 •

edited

Loading