Ancilla vs MZI

Does an ancilla really beat a Mach-Zehnder interferometer under noise? Build a scheme, count it fairly, and see. Everything runs in your browser.

Run

— · 0 seeds

Budget

0 picks one per spare core, up to 4.

Setup

Ancilla on the generator, with coupling, fields and pulses on A.

Input space

0 passive, 1 active; above 1 breaks rule 2.

Coupling αij, S axis i by A axis j
A xA yA z
S x
S y
S z
Static fields
on A
on S
Pulses
Noise

Rates in units of 1/t. The MZI's particles get the same noise.

System

Ancilla

Same noise as the system.

Accounting overrides

Each override breaks one rule; leave them at their defaults for a fair test.

Default N = 2.

Default 1.

Rules

  • Kept 1. Equal particles. The MZI gets all N = 2 particles.
  • Kept 2. Bounded generator. ‖∂H/∂ω‖ₛ = 2 ≤ N = 2.
  • Kept 3. Equal time. Both schemes hold for t.
  • Kept 4. Equal noise on every particle. Every particle, ancilla included, carries the same noise.
  • Always No measurement or feedback during the hold.

Results

Pick a preset or build a setup, then press Run. The MZI baseline comes first, then seeds keep running until you press Stop.

How it works

Pick a preset or build a setup, then press Run. The optimiser finds the best controls for the ancilla scheme and for a plain Mach-Zehnder interferometer, with the same particles, noise and hold time. An R above 1 means the ancilla helps. The Rules panel shows which fairness rules hold.

The model, the rules and the metrics are in the docs.