1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
/// Optimal control via direct shooting methods.
///
/// This module provides no-std, allocation-free implementations of:
///
/// - **ODE solvers** — Euler, RK4, and adaptive RK45 (Fehlberg) integrators
/// used to transcribe the continuous-time dynamics into the shooting NLP.
/// - **Single shooting** — parameterises the entire control sequence `u(t)` and
/// minimises cost via projected gradient descent with Armijo line search.
/// - **Multiple shooting** — introduces auxiliary node states `{s_k}` as
/// decision variables and enforces continuity via a penalty method.
/// - **Pontryagin utilities** — Hamiltonian computation, adjoint co-state
/// dynamics, bang-bang control law, and adjoint-based shooting gradient.
///
/// All implementations are generic over the scalar type `S: ControlScalar`
/// and use const-generic arrays for zero heap allocation.
// ── re-exports ────────────────────────────────────────────────────────────────
pub use MultipleShootingProblem;
pub use ;
pub use ;
pub use ;
// ── error type ────────────────────────────────────────────────────────────────
/// Errors that can arise during optimal control computations.