Lucas T. Brady,Wim van Dam
Abstract:Quantum annealing is guaranteed to find the ground state of optimization problems in the adiabatic limit. Recent work [Phys. Rev. X 6, 031010 (2016)] has found that for some barrier tunneling problems, quantum annealing can be run much faster than is adiabatically required. Specifically, an $n$-qubit optimization problem was presented for which a non-adiabatic, or diabatic, annealing algorithm requires only constant runtime, while an adiabatic annealing algorithm requires a runtime polynomial in $n$.
Here we show that this non-adiabatic speed-up is a direct result of a specific symmetry in the studied problems. In the more general case, no such non-adiabatic speed-up occurs. We furthermore show why the special case achieves this speed-up compared to the general case. We conclude with the observation that the adiabatic annealing algorithm has a necessary and sufficient runtime that is quadratically better than the standard quantum adiabatic condition suggests.
What problem does this paper attempt to address?
This paper attempts to solve a key problem in quantum optimization: under what conditions can the quantum annealing algorithm achieve non - adiabatic acceleration. Specifically, the paper explores whether, under the action of Hamiltonians with certain symmetries, the quantum annealing algorithm can run in non - adiabatic conditions in constant time without having to meet the standard adiabatic conditions, which usually require the running time to grow polynomially with the problem size \(n\).
### Background
Quantum annealing is a method for solving optimization problems. Its core idea is to slowly change the Hamiltonian of the system in the adiabatic limit so that the system finally reaches the ground state of the target Hamiltonian. However, recent research has found that for certain problems, quantum annealing can run in non - adiabatic conditions in constant time, which is much faster than the standard adiabatic conditions.
### Research questions
The paper mainly studies the following questions:
1. **Reasons for non - adiabatic acceleration**: Why can quantum annealing achieve non - adiabatic acceleration under the action of Hamiltonians with certain symmetries?
2. **Behavior in general cases**: In more general cases, does this non - adiabatic acceleration still exist?
3. **Accurate estimation of running time**: How to accurately estimate the optimal running time of the quantum annealing algorithm, especially for different parameters \(\mu\)?
### Main findings
1. **The role of special symmetries**: The paper finds that non - adiabatic acceleration is caused by specific symmetries. Specifically, when the Hamiltonian has a specific form (for example, \(\mu = 1\)), the system can quickly find the ground state in non - adiabatic conditions.
2. **Behavior in general cases**: For slightly more general Hamiltonians (i.e., \(\mu\neq1\)), non - adiabatic acceleration no longer exists. The paper proves this through numerical simulation and theoretical analysis.
3. **Accurate estimation of running time**: The paper proves that in general cases, the optimal running time \(\tau\) of the quantum annealing algorithm has an \(O(\sqrt{n})\) relationship with the problem size \(n\). This is a faster result than the standard adiabatic condition \(O(n)\).
### Conclusions
The main conclusions drawn by the paper are:
- Non - adiabatic acceleration is a special case of Hamiltonians with specific symmetries and is not universal.
- For quantum annealing problems in general cases, the optimal running time \(\tau\) has an \(O(\sqrt{n})\) relationship with the problem size \(n\).
- Achieving non - adiabatic acceleration requires very high time precision, especially in large - scale problems.
These results are of great significance for understanding and optimizing the quantum annealing algorithm, especially in the near - future applications of quantum computers.