Quantum Amplitude Amplification Achieves Optimal Solutions for Combinatorial Problems up to Size 40

Key Takeaways

• Exact oracle parameter formula for linear cost functions.
• Simulations up to 40 qubits show Grover‑like performance.
• First experimental QAA demonstration on IBMQ and IonQ hardware.
• Cost oracles reduce circuit depth versus traditional Grover oracles.
• Diffusion parameter π yields >90% optimal solution probability.

Summary

Researchers from the Air Force Research Laboratory demonstrated that Quantum Amplitude Amplification (QAA) can solve combinatorial optimisation problems up to 40 qubits with near‑optimal efficiency. They derived an exact formula for oracle parameters when the cost function is linear, simplifying circuit design. Simulations confirmed Grover‑like success probabilities, and experiments on IBMQ superconducting and IonQ trapped‑ion devices validated the theory on hardware up to five qubits. The work marks the first experimental use of cost‑oracles in QAA, expanding quantum optimisation toolkits.

Pulse Analysis

Quantum Amplitude Amplification (QAA) is emerging as a powerful alternative to classic Grover search for combinatorial optimisation. By encoding cost functions directly into phase‑shift operators—so‑called cost oracles—researchers bypass the cumbersome multi‑controlled Toffoli gates that have limited Grover’s practicality on near‑term devices. The Air Force Research Laboratory team proved that linear cost functions possess a symmetry enabling an exact, closed‑form calculation of the oracle’s phase parameter, dramatically streamlining circuit synthesis and reducing error‑prone depth.

Extensive simulations spanning 5 to 40 qubits revealed that QAA, when tuned with the derived formula and a fixed diffusion angle of π, achieves success probabilities exceeding 90 % for solutions near the global optimum. This performance closely mirrors the theoretical bound of Grover’s algorithm, confirming that QAA can scale efficiently as problem size grows. The study also highlighted that the algorithm’s iteration count aligns with the square‑root speed‑up expected from quantum search, reinforcing its advantage over classical heuristics for large‑scale binary optimisation tasks such as QUBO.

The theoretical predictions were corroborated on two leading quantum platforms: IBM’s superconducting qubits and IonQ’s trapped‑ion processors. Experiments on up to five qubits, augmented with vendor‑provided noise‑mitigation, showed measured basis‑state probabilities matching the analytical model within statistical error. Demonstrating reliable QAA on heterogeneous hardware not only validates the cost‑oracle approach but also signals readiness for real‑world applications in logistics routing, portfolio optimisation, and materials discovery. Future work will extend the exact parameter formula to non‑linear cost landscapes and explore hybrid QAA‑QAOA strategies, paving the way for more versatile quantum optimisation pipelines.