[General] Release 2.0

[ocots]

Before we release the new 2.0 version of OptimalControl, we need to complete some things.

• Remove prefixes when display the description of a parameter (with a registry), see the describe function. Done here.
• Prefix all CTFlows functions in https://github.com/control-toolbox/CTFlows.jl/blob/c75e770fe9057b5469de077ea24f079b9ce635d2/src/differential_geometry.jl#L612-L675 with CTFlows. Done here.
• Do not reexport types from CTFlows but only CTFlows:

  OptimalControl.jl/src/imports/ctflows.jl

  Line 4 in 0f84eec

  @reexport import CTFlows: Hamiltonian, HamiltonianLift, HamiltonianVectorField

  . Done in Add comprehensive CTFlows signature freezing tests #735.
• Add tests in OptimalControl for the indirect method: flow, resolution but also some Lie computations. Done in Add comprehensive CTFlows signature freezing tests #735 and Add indirect method tests #736.
• Finalise and merge https://github.com/control-toolbox/OptimalControl.jl/pull/718/changes. Done by Refactor canonical tests with separated modelers/solvers #737.
• Finalise and merge feat: support optimal control problems with zero-dimensional control CTParser.jl#246
• Review the documentation for the dim 0 control
• Make a new stable release of CTBase
• Make a new stable release of CTModels
• Make a new stable release of CTSolvers
• Make a new stable release of CTFlows
• Make a new stable release of CTParser
• Make a new stable release of CTDirect which will include 0 dim control support
• Add a word about 0 dim control support in the documentation and some tests
• Add a word about Lie computations in the documentation

Documentation

• replace $\partial(q)$ by $\dot{q}$ (same for $v$):

  OptimalControl.jl/docs/src/example-double-integrator-energy.md

  Lines 52 to 53 in 0f84eec

  	∂(q)(t) == v(t)
  	∂(v)(t) == u(t)

• show the initial guess with a max_iter = 0.
• in the descriptions of the parameters, we have for instance the prefix CTSolvers.Modelers. This should be removed.

[ocots]

Before making the new v2 release, check that all the CTX have a new stable version!

[ocots]

@jbcaillau @PierreMartinon @joseph-gergaud Below this comment is an idea for the future post for the v2. Please modify, tell me if it is too long, if you want to insist on something else, etc.

[ocots]

Title. [ANN] OptimalControl.jl v2.0.0

It is a pleasure to announce the new major release of OptimalControl.jl (v2.0.0). Designed to solve optimal control problems on ODEs, the package features:

• a friendly DSL
• both direct (optimisation) and indirect (aka. shooting) methods
• solving on GPU — now as simple as solve(ocp, :gpu)
• a bunch of examples: tutorials, applications and collection of problems

What's new in v2.0

More solvers, more flexibility

OptimalControl.jl now supports 5 nonlinear optimization solvers (Ipopt, MadNLP, Uno, MadNCL, Knitro) with 12 total solving methods combining different discretization schemes, modelers, and execution backends (CPU/GPU). See the solve manual for the complete list.

GPU solving is now streamlined — just specify :gpu:

sol = solve(ocp, :gpu)  # Automatically uses ExaModels + MadNLP on GPU (with cuDSS)

Learn more in the GPU solving guide.

Control-free problems

A new feature allows solving optimal control problems without control variables — useful for parameter estimation and optimization of constant parameters in dynamical systems. For instance, finding the minimal pulsation $\omega$ for a harmonic oscillator:

ocp = @def begin
    ω ∈ R, variable              # pulsation to optimize
    t ∈ [0, 1], time
    x = (q, v) ∈ R², state
    
    q(0) == 1.0
    v(0) == 0.0
    q(1) == 0.0                  # final condition
    
    ẋ(t) == [v(t), -ω^2 * q(t)]
    
    ω^2 → min   # minimize pulsation
end

sol = solve(ocp)

See the control-free problems example for parameter estimation and optimization applications.

Initial guess with @init macro

Initial guesses can now be constructed with the new @init macro, providing a cleaner and more intuitive syntax:

init = @init ocp begin
    x(t) := [1-t, 0]
    u(t) := 0
end

sol = solve(ocp; init=init)

Check out the initial guess manual for details.

Differential geometry tools documentation

OptimalControl.jl includes differential geometry operators for analyzing Hamiltonian systems and computing singular controls. These tools are now fully documented in v2.0. The @Lie macro provides convenient syntax for Poisson and Lie brackets:

# Define Hamiltonian functions directly
H0(x, p) = p[1] * x[2] + p[2] * x[1]
H1(x, p) = x[1]^2 + p[2]^2

# Compute Poisson brackets with @Lie macro
H01 = @Lie {H0, H1}

# Evaluate at a point
H01([1, 2], [3, 4])

See the differential geometry tools manual and the singular control example.

Direct and indirect methods throughout

All examples in the documentation now systematically feature both direct and indirect approaches, demonstrating the versatility of the package. See for instance the double integrator energy and time minimisation examples.

Comprehensive documentation rewrite

The documentation has been completely rewritten with new manuals covering:

• Solve options and advanced routing
• GPU solving
• Initial guess construction
• Differential geometry tools
• Control-free problems

The friendly DSL you know

Give it a try, and if ever the syntax below does not look familiar enough 🙂...

ocp = @def begin
    t ∈ [0, 1], time
    x ∈ R², state
    u ∈ R, control
    x(0) == [-1, 0]
    x(1) == [0, 0]
    x₂(t) ≤ 1.2
    ẋ(t) == [x₂(t), u(t)]
    0.5∫(u(t)^2) → min
end

... ask your favorite AI to translate the math for you!

About control-toolbox

OptimalControl.jl is part of the control-toolbox ecosystem, which brings together Julia packages for mathematical control and its applications:

• OptimalControl.jl - Tools to model and solve optimal control problems with a friendly DSL
• OptimalControlProblems.jl - Curated collection of benchmark problems
• CTBenchmarks.jl - Benchmarking suite (under construction)

Acknowledgments

We would like to thank all contributors for their work on this package!

Contributing

Contributions are welcome! Whether you want to report bugs, suggest features, or contribute code, feel free to open an issue or start a discussion on GitHub.

Tags: #optimization #control #optimalcontrol #nlp #gpu #julia

[ocots]

@JuliaRegistrator register

Release notes:

Release Notes - OptimalControl.jl v2.0.0

Major version release with complete solve architecture redesign. This release introduces breaking changes from v1.1.6 (last stable release).

Breaking Changes

• Removed functions: direct_transcription → discretize, set_initial_guess → @init macro, build_OCP_solution → ocp_solution
• New solve architecture: methods() now returns 4-tuples (discretizer, modeler, solver, parameter) with required GPU/CPU parameter
• Changed exports: CTBase exceptions and CTFlows types reorganized (see BREAKING.md)

Major Features

Complete Solve Architecture Redesign

Two complementary solving modes:

• Descriptive mode: solve(ocp, :collocation, :adnlp, :ipopt, :cpu) with symbolic strategy specification
• Explicit mode: solve(ocp; discretizer=Collocation(), modeler=ADNLP(), solver=Ipopt()) with typed components
• Partial specification: Auto-completion of missing strategies using first matching method
• Method introspection: methods() lists all 12 available solving methods (10 CPU + 2 GPU)

GPU/CPU Parameter System

• 4th parameter in method tuples for execution backend (:cpu or :gpu)
• Explicit GPU support via :gpu parameter with ExaModels + MadNLP/MadNCL
• 5 solvers available: Ipopt, MadNLP, Uno, MadNCL, Knitro

Advanced Option Routing System

• describe(strategy): Display available options for any strategy
• route_to(strategy=option=>value): Disambiguate shared options between strategies
• bypass(strategy=option=>value): Pass undeclared options to strategies
• Automatic option routing to appropriate components
• Option introspection: options(), option_names(), option_type(), option_description(), option_default()

Initial Guess with @init Macro

• New @init macro for constructing initial guesses
• Alias init for initial_guess keyword argument in solve
• Replaces functional initial guess construction from v1.1.6

Control-Free Problems Support

• Optimal control problems without control variables
• Optimization of constant parameters in dynamical systems
• Augmented Hamiltonian approach: augment=true feature in CTFlows for automatic costate computation
• Simplified flow creation: Flow(ocp) directly creates Hamiltonian flow from control-free problems
• Complete transversality conditions for variable parameters

Additional Discretization Schemes

• Basic schemes: :trapeze, :midpoint, :euler, :euler_implicit
• ADNLP-specific schemes: :gauss_legendre_2, :gauss_legendre_3 (high-order collocation)

Documentation

• Complete rewrite with new solve manual (descriptive/explicit modes)
• Advanced options guide with routing and disambiguation
• GPU solving guide
• Initial guess guide with @init macro
• Differential geometry tools manual
• Control-free problems examples

Dependencies

• CTBase: 0.18.x (was 0.16-0.17)
• CTModels: 0.9.x (was 0.6.x)
• CTDirect: 1.x (was 0.x)
• CTSolvers: 0.4.x (new dependency)
• CTParser: 0.8.x
• CTFlows: 0.8.x

New dependency: CTSolvers.jl handles NLP modeling, solving, and strategy orchestration.

Migration Guide

For users migrating from v1.1.6, please consult BREAKING.md for detailed migration instructions and examples.

Notes

This release consolidates all beta versions (1.2.0-beta through 1.3.1-beta) into a stable 2.0.0 release. The comparison is made against v1.1.6, the last stable release before the architectural redesign.

[JuliaRegistrator]

Registration pull request created: JuliaRegistries/General/152293

Tagging

After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.

This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via:

git tag -a v2.0.0 -m "<description of version>" 2ba73c6e09c7cfd38c2bfd6cdebc5a36fdb5b516
git push origin v2.0.0

[jbcaillau]

thanks @ocots large backlog on my github inbox, couldn't check that post:

• fine, anyway
• regarding contributors, the stats don't give credit to other contributors (e.g. colleagues from MadNLP or so), maybe because they act on CTx repo's but not OC.jl (to be improved, the case being, as OC.jl is only la partie émergée de l'iceberg)

Title. [ANN] OptimalControl.jl v2.0.0

It is a pleasure to announce the new major release of OptimalControl.jl (v2.0.0). Designed to solve optimal control problems on ODEs, the package features [...]

Acknowledgments

We would like to thank all contributors for their work on this package!