Struct AnnealingScheduleOptimizer 

pub struct AnnealingScheduleOptimizer { /* private fields */ }

Quantum Annealing Schedule Optimiser

Learns an optimal annealing schedule s(t) ∈ [0, 1] that minimises the non-adiabatic excitation probability subject to a fixed total annealing time.

The schedule is parameterised as a piecewise-linear function with num_segments control points and optimised via gradient-free COBYLA-style coordinate descent.

Implementations

impl AnnealingScheduleOptimizer

pub fn new(num_segments: usize, total_time: f64, min_gap: f64) -> Self

Create a new schedule optimiser.

pub const fn with_max_iterations(self, n: usize) -> Self

Set the maximum number of optimisation iterations

pub fn linear_schedule(&self) -> Vec<f64>

Generate the default linear schedule: s(t) = t / T.

pub fn roland_cerf_schedule(&self) -> Vec<f64>

Generate a Roland-Cerf optimal schedule.

For a problem with minimum gap Δ_min, the locally-adiabatic (Roland-Cerf) schedule satisfies

ds/dt ∝ Δ(s)²

which spends more time near the avoided crossing. Here we approximate Δ(s) ≈ Δ_min + (1 - Δ_min) · (s - 0.5)² / 0.25 for a parabolic gap.

pub fn optimize_schedule(&self, transverse_field: f64, beta: f64) -> Vec<f64>

Optimise the annealing schedule using a greedy coordinate-descent over the control points.

The objective is to maximise the estimated success probability from a QuantumTunnelingSimulator instantiated with the given parameters.

Returns the optimised schedule control points.