Quantum annealing (QA) is an optimization process for finding the global minimum of a given objective function over a given set of candidate solutions (candidate states), by a process using quantum fluctuations. Quantum annealing is used mainly for problems where the search space is discrete (combinatorial optimization problems) with many local minima, such as finding the ground state of a spin glass or solving QUBO problems, which can encode a wide range of problems like Max-Cut, graph coloring, SAT or the traveling salesman problem. The term "quantum annealing" was first proposed in 1988 by B. Apolloni, N. Cesa Bianchi and D. De Falco as a quantum-inspired classical algorithm. It was formulated in its present form by T. Kadowaki and H. Nishimori (ja) in 1998, though an imaginary-time variant without quantum coherence had been discussed by A. B. Finnila, M. A. Gomez, C. Sebenik and J. D. Doll in 1994.

Quantum annealing starts from a quantum-mechanical superposition of all possible states (candidate states) with equal weights. Then the system evolves following the time-dependent Schrödinger equation, a natural quantum-mechanical evolution of physical systems. The amplitudes of all candidate states keep changing, realizing a quantum parallelism, according to the time-dependent strength of the transverse field, which causes quantum tunneling between states or essentially tunneling through peaks. If the rate of change of the transverse field is slow enough, the system stays close to the ground state of the instantaneous Hamiltonian (also see adiabatic quantum computation). If the rate of change of the transverse field is accelerated, the system may leave the ground state temporarily but produce a higher likelihood of concluding in the ground state of the final problem Hamiltonian, i.e., Diabatic quantum computation. The transverse field is finally switched off, and the system is expected to have reached the ground state of the classical Ising model that corresponds to the solution to the original optimization problem. An experimental demonstration of the success of quantum annealing for random magnets was reported immediately after the initial theoretical proposal. Quantum annealing has also been proven to provide a fast Grover oracle for the square-root speedup in solving many NP-complete problems.

Quantum annealing can be compared to simulated annealing, whose "temperature" parameter plays a similar role to quantum annealing's tunneling field strength. In simulated annealing, the temperature determines the probability of moving to a state of higher "energy" from a single current state. In quantum annealing, the strength of transverse field determines the quantum-mechanical probability to change the amplitudes of all states in parallel. Analytical and numerical evidence suggests that quantum annealing outperforms simulated annealing under certain conditions (see Heim et al and see Yan and Sinitsyn for a fully solvable model of quantum annealing to arbitrary target Hamiltonian and comparison of different computation approaches).

The tunneling field is basically a kinetic energy term that does not commute with the classical potential energy part of the original glass. The whole process can be simulated in a computer using quantum Monte Carlo (or other stochastic technique), and thus obtain a heuristic algorithm for finding the ground state of the classical glass.

In the case of annealing a purely mathematical objective function, one may consider the variables in the problem to be classical degrees of freedom, and the cost functions to be the potential energy function (classical Hamiltonian). Then a suitable term consisting of non-commuting variable(s) (i.e. variables that have non-zero commutator with the variables of the original mathematical problem) has to be introduced artificially in the Hamiltonian to play the role of the tunneling field (kinetic part). Then one may carry out the simulation with the quantum Hamiltonian thus constructed (the original function + non-commuting part) just as described above. Here, there is a choice in selecting the non-commuting term and the efficiency of annealing may depend on that.

It has been demonstrated experimentally as well as theoretically, that quantum annealing can outperform thermal annealing (simulated annealing) in certain cases, especially where the potential energy (cost) landscape consists of very high but thin barriers surrounding shallow local minima. Since thermal transition probabilities (proportional to ${\displaystyle e^{-{\frac {\Delta }{k_{B}T}}}}$, with ${\displaystyle T}$ the temperature and ${\displaystyle k_{B}}$ the Boltzmann constant) depend only on the height ${\displaystyle \Delta }$ of the barriers, for very high barriers, it is extremely difficult for thermal fluctuations to get the system out from such local minima. However, as argued earlier in 1989 by Ray, Chakrabarti & Chakrabarti, the quantum tunneling probability through the same barrier (considered in isolation) depends not only on the height ${\displaystyle \Delta }$ of the barrier, but also on its width ${\displaystyle w}$ and is approximately given by ${\displaystyle e^{-{\frac {{\sqrt {\Delta }}w}{\Gamma }}}}$, where ${\displaystyle \Gamma }$ is the tunneling field. This additional handle through the width ${\displaystyle w}$, in presence of quantum tunneling, can be of major help: If the barriers are thin enough (i.e. ${\displaystyle w\ll {\sqrt {\Delta }}}$), quantum fluctuations can surely bring the system out of the shallow local minima. For an ${\displaystyle N}$-spin glass, the barrier height ${\displaystyle \Delta }$ becomes of order ${\displaystyle N}$. For constant value of ${\displaystyle w}$ one gets ${\displaystyle \tau }$ proportional to ${\displaystyle e^{\sqrt {N}}}$ for the annealing time (instead of ${\displaystyle \tau }$ proportional to ${\displaystyle e^{N}}$ for thermal annealing), while ${\displaystyle \tau }$ can even become ${\displaystyle N}$-independent for cases where ${\displaystyle w}$ decreases as ${\displaystyle 1/{\sqrt {N}}}$.

It is speculated that in a quantum computer, such simulations would be much more efficient and exact than that done in a classical computer, because it can perform the tunneling directly, rather than needing to add it by hand. Moreover, it may be able to do this without the tight error controls needed to harness the quantum entanglement used in more traditional quantum algorithms. Some confirmation of this is found in exactly solvable models.

Timeline of ideas related to quantum annealing in Ising spin glasses: