Skip to content

[Tracking] Per-modeling lists of problems, drift-proof tests and cron #310

Open
13 of 16 issues completed
Open

Description

@ocots

Goal

Reorganise OptimalControlProblems.jl around four ideas:

  1. Each modeling has its own list of problems, discovered automatically from the files: a problem may exist in OptimalControl without a JuMP version. No list of problems is written by hand anywhere.
  2. Every problem keeps a comparison: another modeling (automatic OptimalControl / JuMP tests on the intersection), or a sourced reference_objective.
  3. Drift is caught by tests: fast consistency tests on every PR, slow tests in a weekly cron (as in CTFlows.jl).
  4. The collection grows, including from SciML's OptimalControlBenchmarks (Corleone), with clean documentation of all this.

This builds on the port to OptimalControl v2 (#283) and the JuMP schemes (#284), both tracked in control-toolbox/CTBenchmarks.jl#200 for the benchmark paper. Roadmap: #257.

Work, in order

Step 0 brings the repository infrastructure in line with the other packages; steps 1 to 3 are the base; steps 4 to 7 can then go in parallel.

  1. Repository infrastructure: [Chore] Harmonize repository metadata and ct-conventions integration with CTBase #313 (harmonization with ct-conventions: AGENTS.md, labels, templates, lint), then [Chore] Release and CI infrastructure: CHANGELOG, BREAKING, release workflows, label-gated CI #312 (release infrastructure). Agents working on the other sub-issues need AGENTS.md and the pinned conventions.
  2. Per-modeling lists and consistency tests: [Dev] Per-modeling lists of problems, discovered automatically, with a status per modeling #303, [Tests] Test layout per environment and consistency tests #304 (with [Doc] Typo in the docstrings #172, [Doc] Examples are not displayed correctly in the REPL #173 under [Dev] Per-modeling lists of problems, discovered automatically, with a status per modeling #303).
  3. Port to OptimalControl v2 and new problem interface: [Paper] Port to OptimalControl v2 and new problem interface #283 (with [Dev] Exa and problems #136). The problem returns the continuous problem and its initial guess; modeler, options, scheme and CPU/GPU are chosen in solve, so this package no longer passes modeler options.
  4. Weekly cron: [Tests] Crons per scope for the slow tests, with reports #305.
  5. JuMP with Euler, trapezoid and midpoint, factorised in ext/JuMPModels/schemes.jl, and NLP equivalence tests against ADNLP and Exa: [Paper] JuMP models for Euler, trapeze and midpoint, and NLP equivalence tests #284 (with [General] Test comparaison NLP model #166).
  6. Metadata and shared initial guess: [Metadata] Computed features, reference objective, shared initial guess #306 (with [Doc] Intial guess #151); measure, then keep or drop the _s variants: Description and problems #174.
  7. Plots in a Plots extension, shared by OptimalControl solutions and JuMP models: [General] Plots #175, [General] Visualisation #134; documentation: [Doc] Rewrite the documentation: getting started, user guide, problems, developers #307 (with [Doc] review of examples published on the website #229).
  8. Corleone's OptimalControlBenchmarks: [Problems] Proof of concept: a ModelingToolkit modeling, solved with Corleone, with our own bryson_denham #308.
  9. Precompile the problems, so that users and tests do not recompile them (later, not a priority): [Dev] Precompile the problems so that users and tests do not recompile them every time #373.
  10. Restore the full CI matrix (Julia 1.10, macOS, Windows) at the end of the reorganisation: Restore the full CI matrix (Julia 1.10, macOS, Windows) after the reorganisation #344.

Decisions

What each modeling returns

Rule: each modeling returns the most abstract object it can represent, and takes at retrieval only what it needs to build it.

Modeling Returned object Arguments at retrieval
OptimalControl continuous problem: struct (name, ocp, init, metadata) parameters
JuMP NLP, necessarily discretized grid_size, scheme, JuMP.Model arguments
Corleone (PoC) MTK system and its grids (path constraints are sampled in the constructor) control_grid, shooting_grid, constraint_grid

grid_size and scheme in JuMP use exactly the names and values of OptimalControl.Collocation (:euler, :euler_implicit, :trapeze, :midpoint), so that tests build both NLPs identically.

Checks between modelings

How to work on this

Each sub-issue is closed by one or a few PRs. Tick the decisions above when taken, and record them in the relevant sub-issue.

Activity

  1. ocots commented on Nov 4, 2026

    @ocots
    MemberAuthor

    Decisions taken while detailing #303:

    • The _s variants become a modeling of their own, OptimalControlScalarizedBackend, maintained like the others (no more p_s functions).
    • Each problem has a status per modeling in its metadata: :draft or :valid. A problem is provided (problems(backend)) only where it is :valid; :valid in a second modeling requires consistency with the first. Drafts stay in the repository; calling one emits a warning. The status board (problem × modeling) is generated in the documentation, and the cron reports drafts ready to be validated.
    • Docstrings are generated from one template: short, with a usage example per modeling and a link to the online page.
    • Tests are split into environments (core, one per modeling, comparison, Corleone); the cron is a mode of each. Moved to [Tests] Test layout per environment and consistency tests #304.
    • No release for now: work on main.
    • Guiding principle: everything that can be computed is generated or checked automatically, never maintained by hand.
  2. 6 remaining items

  3. ocots commented on Nov 4, 2026

    @ocots
    MemberAuthor

    Update: #303 is closed (PR #320, with #319 = #304 steps 1–2 and 6, #321, #322). New follow-up #323 (test of the automatic discovery and of the precompilation cache), difficulty ★★, can be taken by an agent once #304 is done (it uses the layout test/suite/core/). Julia matrix is now 1.10 and 1.13. The label run ci github-runner now triggers the CI by itself.

  4. ocots commented on Nov 4, 2026

    @ocots
    MemberAuthor

    Correction to my previous update: #323 can be taken now — it only needs the layout test/suite/core/, already on main (#319). Starting notes are in #323 (comment) (branch from main; respect the metadata schema and the no-hand-written-list test of #324).

  5. ocots commented on Nov 4, 2026

    @ocots
    MemberAuthor

    Follow-ups opened since the guide (so that they are not forgotten). They extend the table of the guide; same rules (claim, worktree, draft PR).

    # Issue Difficulty Depends on When
    #323 Test the automatic discovery of problems and the precompilation cache ★★ the layout of test/suite/core/ (on main) now; PR #325 is in review
    #327 Test only the problems a pull request changes (CI job on top of the test argument problem=a,b) ★★ #304 (PR #324 adds problem=) after #324 is merged
    #326 Align the costate and dual signs of the JuMP models with OptimalControl (SGN1–SGN6) ★★ #284 (JuMP schemes) and the ecosystem discussion on multipliers (control-toolbox discussion #68), which it waits for after #284; then remove the -p(t) workaround of the docs (#307, #314) and add costate comparisons to comparison
    CTActions#133 Release gate: refuse scope labels on a release PR ★★ — before the next release of this package (after #283); until then remove every test … label from a release PR
    CTActions#134 Cancel superseded runs of a PR in the reusable workflows ★★ — any time; in CTActions, not here

    Notes recorded elsewhere, to apply when taking the issue concerned:

  6. ocots commented on Nov 5, 2026

    @ocots
    MemberAuthor

    Table of the sub-issues, updated 2026-11-05

    This table replaces the two tables of the guide (section 1) and of the follow-ups comment. The rules for agents are unchanged and are in the guide: read it first (choosing and claiming an issue, one worktree per issue, draft PR, narration, stop and report, never merge, never edit a description). An issue can be started only when every issue it depends on is closed (its PR merged).

    Difficulty: ★ mechanical · ★★ several files, a few choices · ★★★ design inside a detailed specification · ★★★★ delicate (exact equivalence, new infrastructure, cross-repository effects).

    Done

    Issue Result
    #313 Harmonize with ct-conventions closed
    #312 Release and CI infrastructure closed
    #303 Per-modeling lists and statuses closed
    #304 Test layout, statuses, consistency tests PRs #319, #320, #324 merged; the issue is open until the maintainer closes it. What it leaves to others is in its closing comment and in notes on #283, #284, #305, #306, #311

    To do, in order

    # Issue Difficulty Depends on Status Can run in parallel with
    1 #283 theme: port to OptimalControl v2, in three parts below (spec in the comments of #283) ★★★★ #303, #304, #312 (done) in progress #305 and everything not after it
    1a #328 part 1: Problem type, continuous problems, CPU tests ★★★★ done taken, draft PR #331 #305
    1b #329 part 2: CTModels.Solution from a solved JuMP model ★★★ #328 waits for #328 #330
    1c #330 part 3: GPU tests for the OptimalControl modelings ★★★ #328 waits for #328 #329
    1d #332 replace LocalTestRunner by CTBase.run_tests ★★ #328 waits for #328 (it says "#331 merged": #331 is still a draft) #329, #330
    2 #305 Crons per scope, reports, bug issues ★★★ #304 (done) free #283
    3 #284 JuMP schemes and NLP equivalence (spec in a comment) ★★★★ #283 waits for #283 #306, #311, #305
    4 #306 Computed features, reference objective, shared initial guess (+ local solutions, robustness of the initial guess: see its note) ★★★ #283 waits for #283 #284, #311, #305
    5 #311 Migration to DocumenterVitepress ★★ #283 waits for #283 #284, #306, #305
    6 #314 Results pipeline and gallery ★★★★ #311, #306 waits #307
    7 #307 Documentation rewrite ★★★ #311 waits #314
    8 #175 Plot OptimalControl and JuMP solutions together ★ #283 (#329 for the JuMP solution), #284 waits #307, #314
    9 #134 Problem-specific visualisation (low priority) ★★ #283 waits anything after #283
    10 #308 ModelingToolkit modeling solved with Corleone (proof of concept) ★★★★ #303, #304, #306, #307, #314 waits (last) —

    Independent follow-ups (any time, unless stated)

    Issue What Difficulty When
    #323 Test the automatic discovery of problems and the precompilation cache ★★ taken, PR #325 in review; to rebase on main now that #324 is merged
    #327 Test only the problems a pull request changes (CI job on top of problem=a,b, available since #324) ★★ free
    #326 Align the costate and dual signs of the JuMP models with OptimalControl (SGN1–SGN6) ★★ after #284 and the ecosystem discussion on multipliers (control-toolbox discussion #68); then remove the -p(t) workaround of the docs (#307, #314)
    CTActions#133 Release gate: refuse scope labels on a release PR ★★ before the next release of this package; until then remove every test … label from a release PR
    CTActions#134 Cancel superseded runs of a PR in the reusable workflows ★★ in CTActions, not here

    Not scheduled

    Notes recorded elsewhere, to apply when taking the issue concerned

  7. ocots commented on Nov 5, 2026

    @ocots
    MemberAuthor

    CI matrix reduced during the reorganisation (decision of the maintainer).

    Rules of the first comment of this issue are unchanged.

  8. ocots commented on Nov 5, 2026

    @ocots
    MemberAuthor

    Status, 2026-11-05: items outside the steps above

    Open and not part of steps 0–7, to keep in view:

    Step 2 (#283): part 1 is PR #331 (draft, CI running), then #329 and #330.

    🤖 Generated with Claude Code

  9. ocots commented on Nov 5, 2026

    @ocots
    MemberAuthor

    Table of the sub-issues, updated 2026-11-05 (evening)

    This table replaces the table of this morning and the status of 11:02. The rules for agents are unchanged and are in the guide: read it first (choosing and claiming an issue, one worktree per issue, draft PR, narration, stop and report, never merge, never edit a description). An issue can be started only when every issue it depends on is closed.

    Difficulty: ★ mechanical · ★★ several files, a few choices · ★★★ design inside a detailed specification · ★★★★ delicate (exact equivalence, new infrastructure, cross-repository effects).

    Done

    Step Issue Delivered by
    0 #313 harmonize with ct-conventions; #312 release and CI infrastructure closed
    1 #303 per-modeling lists and statuses; #304 test layout and consistency tests #319, #320, #324
    1 #323 test of the discovery and of the precompilation cache #325
    2 #283 port to OptimalControl v2: #328 problem type (#331), #329 CTModels.Solution from JuMP (#353), #330 GPU tests (#354, with #357), #332 CTBase.run_tests (#349) closed
    3 #305 crons per scope and reports #335, #346
    CI #327 test only the changed problems (#334); #341 report what a run tested (#355) closed
    Tests #360 report convergence instead of testing it: smoke tests (small grid, few iterations), equivalence always hard, convergence off by default, on in the crons, on demand with the argument or label test convergence #363, #365
    Outside CTActions#134 cancel superseded runs closed

    To do, by independent package

    Packages A, B and C can run in parallel; D waits for B; E is last.

    A. JuMP schemes and NLP equivalence — theme #284 (spec in its comments)

    # Issue Difficulty Depends on Status
    A1 #368 write the trapezoidal scheme once in schemes.jl ★★★ — taken, draft PR #374
    A2 #369 Euler, implicit Euler, midpoint ★★★ #368 waits
    A3 #370 ordering of variables and constraints (interleaved or blocks) ★★ #368 waits
    A4 #371 NLP equivalence tests JuMP / ADNLP / Exa ★★★★ #369, #370 waits
    A5 #372 explain the ducted_fan Hessian discrepancy (#166) ★★★ — free
    A6 #326 costate and dual signs of the JuMP models ★★ #284, discussion #68 on multipliers waits

    B. Metadata and references — #306

    # Issue Difficulty Depends on Status
    B1 #306 computed features, reference objective, shared initial guess (+ local solutions and robustness of the initial guess, see the notes below) ★★★ #283 (done) free; the shared initial guess touches ext/JuMPModels/<problem>.jl, rewritten by #368: coordinate, or do it after #368

    Once references exist, the convergence report (#360) shows the gap to them; it is empty until then.

    C. Documentation

    # Issue Difficulty Depends on Status
    C1 #311 migration to DocumenterVitepress ★★ #283 (done) taken (claim of 2026-11-05)
    C2 #307 documentation rewrite ★★★ #311 waits
    C3 #134 problem-specific visualisation (low priority) ★★ #283 (done) free

    D. Results and plots (after B and C)

    # Issue Difficulty Depends on Status
    D1 #314 results pipeline and gallery ★★★★ #311, #306 waits
    D2 #175 plot OptimalControl and JuMP solutions together ★ #284 waits

    E. Last

    # Issue Difficulty Depends on Status
    E1 #308 ModelingToolkit modeling solved with Corleone (proof of concept) ★★★★ #306, #307, #314 waits
    E2 #344 restore the full CI matrix (Julia 1.10; Ubuntu, macOS, Windows) and fix what it shows. Do not close #310 before it. ★★ everything else waits (last step before the release)

    Independent, any time

    Issue What Difficulty When
    #367 cron report: moonlander (OptimalControl, Windows) no longer converges (ITERATION_LIMIT) ★★ free; look at the run, then decide: expected or regression
    #373 precompile the problems so that users and tests do not recompile them (measurements in its comments: almost all the test time is compilation) ★★★★ later, not a priority; measure with SnoopCompile first
    CTActions#133 release gate: refuse scope labels on a release PR ★★ before the next release; until then remove every test … label from a release PR

    Not scheduled

    Notes to apply when taking the issue concerned

    🤖 Generated with Claude Code

  10. ocots commented on Nov 7, 2026

    @ocots
    MemberAuthor

    Update, 2026-11-07: cron reports triaged, new issue #409.

    It joins package B of the table of 2026-11-05 (evening):

    # Issue Difficulty Depends on Status
    B2 #409 typed references with how they were obtained ★★★ #375, #378, #379 (done) taken

    Coordination: #408 (docs) reads the reference files; #409 only adds fields. #314 keeps the trajectories, plots and full environments of the runs.

    🤖 Generated with Claude Code

  11. ocots commented on Nov 7, 2026

    @ocots
    MemberAuthor

    Table of the sub-issues, updated 2026-11-07 (afternoon)

    This table replaces the table of 2026-11-05 (evening) and the update of 2026-11-07.

    Before taking an issue, read the guide for agents. It covers choosing and claiming an issue, one worktree per issue, the plan before code, draft PRs, narration, stopping and reporting, never merging and never editing a description. Its rules are unchanged. An issue can be started only when every issue it depends on is closed.

    Difficulty: ★ mechanical · ★★ several files, a few choices · ★★★ design inside a detailed specification · ★★★★ delicate (exact equivalence, new infrastructure, cross-repository effects).

    Free: to take, by theme

    Code structure

    Issue What Difficulty Depends on Note
    #413 split src/ into submodules (MOD1, MOD2), keeping the public exports (decision in the issue) ★★★ #408 (merged) Do it before part 2 of #409, whose types go into References/.

    JuMP models

    Issue What Difficulty Depends on Note
    #326 align the costate and dual signs of the JuMP models with OptimalControl (with #170) ★★ #284 (done); discussion #68 on multipliers —

    Documentation

    Issue What Difficulty Depends on Note
    #407 local InterLinks inventories (SITE4) ★ — The SVG plots are done in #408.
    #134 problem-specific visualisation ★★ — Low priority.

    Results and plots

    Issue What Difficulty Depends on Note
    #314 results pipeline: plots, freshness badge, Reproducibility block per problem, gallery ★★★★ #311, #306 (done) The problem pages of #408 already show the results of the reference files. #314 keeps the trajectories and the full environments of the runs; #409 only adds fields to the reference files: coordinate with it.
    #175 plot OptimalControl and JuMP solutions together ★ #284 (done) With #314.

    Waiting

    Issue What Difficulty Waits for
    #308 ModelingToolkit modeling solved with Corleone (proof of concept); one documentation adapter to add (docs/modelings.jl, #408) ★★★★ #307, #314 (#306 done)
    #391 theme: one source of truth per problem. Point 1 is done (#377, #379), and #399 (#400). — Its last two parts wait for #308 and for control-toolbox/CTDirect.jl#673.
    #229 review of the examples of the website ★ #307
    #344 restore the full CI matrix (Julia 1.10; Ubuntu, macOS, Windows) and fix what it shows. Do not close #310 before it. ★★ everything else: the last step before the release

    Independent, any time

    Issue What Difficulty When
    #398 remove the MadNLP barrier bypass of the robustness tests ★ once the test environments use CTSolvers ≥ 0.7.1 (OptimalControl and CTDirect still require 0.5)
    CTActions#133 release gate: refuse scope labels that restrict the tests of a release PR ★★ before the next release; until then, remove every test … label from a release PR
    control-toolbox/CTModels.jl#515 build_solution from model information — outside this repository, independent

    Not scheduled

    Notes to apply when taking an issue

    In progress, or almost done

    Issue What Status
    #409 reference solutions with how they were obtained: AbstractReference, AnalyticReference, NumericalReference, qualified API (★★★) taken. Part 1 is draft PR #412 (test/ only, independent of #413). Part 2 comes after #413 and after crons on main.
    #410 a PR without the preview label announces a deployment that never happens (★★) taken, PR #411
    #307 rewrite the documentation Delivered by #408. To close by the maintainer after review; #229 keeps the note on the MKL warning (macOS ARM, upstream).

    Done

    Theme Issues Delivered by
    Infrastructure #313 conventions; #312 release and CI infrastructure closed
    Lists and statuses #303 per-modeling lists and statuses; #304 test layout and consistency; #323 discovery and precompilation cache #319, #320, #324, #325
    OptimalControl v2 #283: #328 problem type, #329 solution from JuMP, #330 GPU tests, #332 CTBase.run_tests #331, #353, #354, #349
    Crons #305 crons per scope and reports; #367, #381 reports triaged #335, #346
    Tests #360 report convergence instead of testing it #363, #365
    CI #327 test only the changed problems; #341 report what a run tested; #405 test what a PR changes, labels only add #334, #355, #406
    JuMP schemes and NLP equivalence (#284) #368 trapezoidal scheme written once; #369 Euler, implicit Euler, midpoint; #370 orderings; #371 NLP equivalence JuMP / ADNLP / Exa (with #166); #372 ducted_fan Hessian; #399 scalarized models; #401 schemes compared on problems chosen by their features #374, #383, #390, #396, #400, #402
    Metadata (#306) #375 reference objectives; #376 features computed from the model; #377 shared initial guess (with #151); #378 known solutions and robustness; #379 analytic solution, sources (tags and Corleone dropped) #380, #392, #395, #393 and #394, #403
    Documentation #311 DocumenterVitepress; #307 rewrite (to close after review); #172, #173 #382, #408
    Not planned #373 precompile the problems: measured and prototyped (#404); too costly for users, and the CI time is reduced by #405 instead closed
    By the maintainer #174, #136 closed
    Outside CTActions#134 cancel superseded runs; CTActions#141 GPU artifact name (#142); breakage: nothing to test is no longer an error (CTActions#143); CTSolvers.jl#321 mu_init and barrier reach MadNLP (#322) merged

    🤖 Generated with Claude Code

  12. ocots commented on Nov 11, 2026

    @ocots
    MemberAuthor

    Written by Claude (Claude Code) at the request of ocots.

    Table of the sub-issues, updated 2026-11-11

    This table replaces the table of 2026-11-07 (afternoon), which is kept as a trace. Most of what it listed as free or waiting has been delivered since.

    Before taking an issue, read the guide for agents. It covers choosing and claiming an issue, one worktree per issue, the plan before code, draft PRs, narration, stopping and reporting, never merging and never editing a description. Its rules are unchanged. An issue can be started only when every issue it depends on is closed.

    Difficulty: ★ mechanical · ★★ several files, a few choices · ★★★ design inside a detailed specification · ★★★★ delicate (exact equivalence, new infrastructure, cross-repository effects).

    Where we are

    Steps 0 to 6 of the plan above are delivered, and step 7 (#308, ModelingToolkit with Corleone) is merged and waits for its first cron result and a decision. What is left is the review of the result (#449), a few follow-ups, and the CI matrix (#344), the last step before the release.

    Open: to take or to follow

    JuMP models

    Issue What Difficulty Depends on Note
    #430 expose the duals of the JuMP path, boundary and box constraints, aligned with rules/signs.md ★★★ #326 (costate done, #431) Follow-up of #431; the carrier-aware signals are in control-toolbox/CTModels.jl#518.
    #326 align the costate and dual signs of the JuMP models with OptimalControl — — #431 delivered the costate (-JuMP.dual, stamped at the step midpoints, compared with OptimalControl in test_solution, docs workaround removed). Still open: to close by the maintainer, or to say what remains; the other duals are #430.

    Results and plots

    Issue What Difficulty Depends on Note
    #175 plot OptimalControl and JuMP solutions together ★ #326 (costate sign, done) solution(model, pb) exists (#353) and the costates now have the same sign. Problems available in JuMP only still wait for control-toolbox/CTModels.jl#515.

    Tests

    Issue What Difficulty Depends on Note
    #445 share the GPU smoke test between OptimalControl and OptimalControlScalarized ★ #441 (done, #446) The shared runners now exist: test/common/optimalcontrol_tests.jl over test/common/modeling_tests.jl.
    #398 remove the MadNLP barrier bypass of the robustness tests ★ CTSolvers ≥ 0.7.1 in the test environments OptimalControl and CTDirect still require 0.5.

    Review

    Issue What Difficulty Note
    #449 review of the rewritten repository, by theme (T1 to T8) ★★ to ★★★ One theme at a time, findings as comments on the issue, a small PR per theme. No theme is ticked yet.

    Waiting

    Issue What Difficulty Waits for
    #308 ModelingToolkit modeling solved with Corleone (proof of concept) ★★★★ The first cron run on main (launched on 2026-11-11, run 43061056587) and the decision of the maintainer: write more problems this way, or not. See below.
    #391 theme: one source of truth per problem. Point 1 is done (#377, #379), and #399 (#400). — Its last two parts wait for the decision of #308 and for control-toolbox/CTDirect.jl#673.
    #287 new problems for the paper: analytic solutions, large state dimension, roadmap ★★★ Not scheduled; do not start it without the maintainer.
    #344 restore the full CI matrix (Julia 1.10; Ubuntu, macOS, Windows) and fix what it shows. Do not close #310 before it. ★★ everything else: the last step before the release

    State of #308

    Delivered by #440 (#437), #442 (#438) and #444 (#439):

    • ModelingToolkitBackend and the ModelingToolkitModels extension; four continuous problems, bryson_denham, vanderpol, robbins and jackson (a Lagrange cost, state bounds with boundary conditions, a Mayer cost with state and control bounds).
    • The solve with Corleone and Ipopt (test/common/solve_modelingtoolkit.jl): single shooting, controls piecewise constant, path constraints imposed at the nodes of the grid, controls started from the initial guess of the problem, the violation of the constraints computed, the trajectory read from the problem solved.
    • The ModelingToolkit environment, opt-in (label test ModelingToolkit backend), and a weekly cron (Saturday 04:00 UTC, Ubuntu and Windows). The tests follow the three kinds of [Tests] Report convergence instead of testing it; keep hard checks for breakage and model equivalence #360: the Problem API and the smoke solves are shared with the OptimalControl modelings (test/common/modeling_tests.jl), and the equivalence with the OptimalControl and the scalarized models is checked at the continuous level, without a solve, by suite/ModelingToolkit/test_modelings (the System is read as the data of a JuMP problem and compared by Continuous.compare, Rationalize JuMP and comparison test suites: separate generic transcription unit tests from per-problem modeling checks #441).
    • A fix in tools/ci/scope.py: an opt-in environment asked for by a label was lost when a shared file made the CI run everything.

    Differences with the specification of the issue, to take into account in the decision (the reasons are in this comment of #308):

    • No test_reference_solution and no cross-evaluation with Corleone's objective. Convergence is reported, not tested ([Tests] Report convergence instead of testing it; keep hard checks for breakage and model equivalence #360), and the equivalence is checked at the continuous level, which catches a wrong model on every pull request (each change of a parameter of the ModelingToolkit model fails exactly the matching check).
    • The documentation snippet stops at the model. The shooting choices (the grids, the controls piecewise constant) are not part of the model, so the page does not show a solve with Corleone; the results of the cron are drawn as for the other modelings. checked_against is nothing, since the comparison runs in the opt-in environment only.
    • The gap of the report is to the analytic objective, since a shooting solve has no reference objective of its own discretization; the reference_objective of [Metadata] Reference objectives per scheme, computed by a script #375 is the trapezoidal one.
    • Limits of the continuous comparison: a problem with a variable (a free time) is not read yet; in SciML's catalogue this is a time-scaling parameter. The control grid and the constraint grid are the same, single shooting.
    • SciML's OptimalControlBenchmarks has about thirty problems, among them most of ours (bryson_denham, cart_pendulum, dielectrophoretic_particle, double_oscillator, ducted_fan, jackson, mountain_car, robbins, van_der_pol, hang_glider, goddarts_rocket, ...), a source of inspiration for the problems to port.

    Independent, any time

    Issue What Difficulty When
    CTActions#133 release gate: refuse scope labels that restrict the tests of a release PR ★★ before the next release; until then, remove every test … label from a release PR
    control-toolbox/CTModels.jl#515 build_solution from model information — outside this repository, independent
    control-toolbox/CTModels.jl#518 carrier-aware signals — outside this repository; needed for #430

    The cron summary issues (#435 cron-failure, #436 cron-convergence) are generated: triage them as described in .github/workflows/README.md.

    Done since the table of 2026-11-07

    Theme Issues Delivered by
    ModelingToolkit (#308) #437 backend and extension; #438 Corleone solve helper and test suite; #439 cron, trajectories, documentation #440, #442, #444
    Results and documentation #314 results pipeline and gallery of the problems; #134 visualisation; #307 rewrite of the documentation; #407 InterLinks (dropped, #416); #229 review of the examples; #410 deploy decision #417, #419, #420, #422, #408, #411
    Code structure #413 src/ split into submodules #421
    Metadata #409 references with how they were obtained (typed values); #286 problem metadata; #285 the _s variants measured and cleaned #412, #423
    JuMP #326 costate sign (the duals: #430 open) #431
    Tests #441 JuMP and comparison suites rationalized: the generic transcription tested once, every problem compared with the others at the continuous level, the NLP of every problem compared with CTDirect's in a weekly cron (comparison) #446
    Crons and CI #432 readable summary issues for the crons; label dispatch (CTActions#149); statuses permission of the called workflows; ct-conventions pinned at v0.12.0 #433, #447, #448, #450
    Roadmap #257, #179 closed

    The items of the earlier tables (infrastructure, lists and statuses, OptimalControl v2, crons, tests, CI, JuMP schemes, metadata) are unchanged: see the table of 2026-11-07.

    Notes to apply when taking an issue

    🤖 Generated with Claude Code

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions