[Roadmap] Update OptimalControlProblems for v2 and add new problems

[ocots]

Objective

Update OptimalControlProblems.jl for OptimalControl v2 compatibility and add new benchmark problems from the curated list of sources below.

Context

OptimalControlProblems.jl provides a curated collection of benchmark optimal control problems. It needs to be updated for the v2 API, and new problems need to be added. This issue consolidates and supersedes issue #92.

Problem sources

A non-exhaustive list of places to find benchmark optimal control problems:

• ICLOCS — Imperial College London Optimal Control Software, example library
• COPS — Constrained Optimization Problem Set (Argonne)
• PSOPT — PSOPT example list; see also the PSOPT Application Examples Document and the source examples
• MINTOC — Mintoc optimal control toolkit examples
• DYMOS — Dymos (OpenMDAO) example library
• BOCOP — BOCOP examples document
• Betts — Practical Methods for Optimal Control and Estimation Using Nonlinear Programming, second edition (Betts)

Specific problems to add

• MRI — from issue [Dev] Add examples #92
• Kepler — from issue [Dev] Add examples #92
• Swimmer — see swimmer.jl and swimmer2.jl (scalarised version) in CTDirect test problems
• Problems from the PSOPT examples that overlap with existing OptimalControlProblems could be updated/cross-referenced

Why

A rich problem library is essential for benchmarking, testing, and tutorials. Keeping it up to date with the latest API ensures it remains useful for the community. The benchmarking paper (see roadmap) will rely on this collection.

References

• OptimalControlProblems.jl
• Issue #92 (superseded, to be closed)
• Roadmap comment