Quantum Zeno effect versus adiabatic quantum computing and quantum annealing
Naser Ahmadiniaz, Dennis Kraft, Gernot Schaller, Ralf Schützhold · 2025
Read Quantum Zeno effect versus adiabatic quantum computing and quantum annealing by Naser Ahmadiniaz, Dennis Kraft, Gernot Schaller, Ralf Schützhold (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-04057.txt.
Introduction
Naser Ahmadiniaz and three co-authors examine a clash between quantum computing and its environment in a 2025 physics paper. Adiabatic quantum computing changes a system very slowly so it can settle into an answer. Quantum annealing uses a similar slow change to solve hard problems. Both rely on smooth jumps between energy levels at narrow crossing points, like lanes merging on a road.
Surrounding noise constantly measures the system, the authors show, and that watching freezes change under the quantum Zeno effect. For a slow version of Grover's search method, the freeze badly limits any speed gain. Gentler state changes or echo tricks that undo noise might ease the problem, they suggest.
Catalog record
| CREATOR | Naser Ahmadiniaz, Dennis Kraft, Gernot Schaller, Ralf Schützhold |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-04057
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quantum Zeno effect versus adiabatic quantum computing and quantum annealing Naser Ahmadiniaz, Dennis Kraft, Gernot Schaller, Ralf Schützhold Abstract For the adiabatic version of Grover’s quantum search algorithm as proposed by Roland and Cerf, we study the impact of decoherence caused by a rather general coupling to some environment. For quite generic conditions, we find that the quantum Zeno effect poses strong limitations on the performance (quantum speed-up) since the environment effectively measures the state of the system permanently and thereby inhibits or slows down quantum transitions. Generalizing our results, we find that analogous restrictions should apply universally to adiabatic quantum algorithms and quantum annealing schemes which are based on isolated Landau-Zener type transitions at avoided level crossings (similar to first-order phase transitions). As a possible resort, more gradual changes of the quantum state (as in second-order phase transitions) or suitable error-correcting schemes such as the spin-echo method may alleviate this problem. I Introduction I.1 Quantum Zeno Effect The Greek philosopher Zeno (or Zenon) formulated several apparent paradoxes illuminating the tension between the two concepts of being in motion versus being at a certain position at a given time. As a famous example, imagine a race between the fast Achilles and a slow tortoise where both start at the same time but the tortoise gets a head start. Then, before being able to overtake the tortoise, Achilles must first reach the place where the tortoise started – but, during that time, the tortoise already moved away and covered an additional distance. Thus, Achilles needs some more time to cover this distance too, which allows the tortoise to move even further, and so on. This leads to a never ending sequence such that one might expect that Achilles should never be able to overtake the tortoise. Intriguingly, this apparent paradox resurges in quantum theory misra1977a ; itano1990a ; hackenbroich1998a ; fischer2001a ; gurvitz2003a ; streed2006a ; bernu2008a ; slichter2016a ; blumenthal2022a ; ahmadiniaz2022a where frequent measurements of the state of a system (corresponding to the concept of position) may slow down or even inhibit the quantum evolution in the form of a transition to another state (corresponding to the concept of motion). As a simple example, let us consider the unitary evolution of a spin- 1/2 system \left|\psi(t)\right> which rotates \left|\psi(t)\right>=\cos(\Omega t)\left|\uparrow\right>+\sin(\Omega t)\left|\downarrow\right> from the initial state \left|\uparrow\right> to the final state \left|\downarrow\right> . Measuring the state of the system (in the \left|\uparrow\right>,\left|\downarrow\right> basis) after a short time \Delta t\ll 1/\Omega projects it to one of the basis states with the probabilities P_{\uparrow}=\cos^{2}(\Omega\Delta t) or P_{\downarrow}=\sin^{2}(\Omega\Delta t)\approx(\Omega\Delta t)^{2} . Repeating this measurement N times with N\Omega\Delta t=\pi/2 then yields the total transition probability 1-\cos^{2N}(\Omega\Delta t) which scales as N(\Omega\Delta t)^{2}\ll 1 , i.e., the transition rate is strongly reduced from \,{\cal O}(\Omega) to \,{\cal O}(\Omega^{2}\Delta t) . As one way to understand this reduction, we have to sum small probabilities (in the case of measurements) instead of small amplitudes (for unitary evolution). Besides the projective measurements discussed above, weak measurements aharonov2014a ; kastner2017a can have a similar effect breuer2007 ; wiseman2010 – provided that the system dynamics is slow enough. In order to illustrate this point, let us consider the standard Lindblad master equation ( \hbar=1 ) \displaystyle\frac{d\hat{\rho}}{dt}=-i\left[\hat{H},\hat{\rho}\right]+\hat{L}\hat{\rho}\hat{L}^{\dagger}-\frac{1}{2}\left\{\hat{L}^{\dagger}\hat{L},\hat{\rho}\right\}\,, (1) where the system Hamiltonian \hat{H} generates the unitary time evolution of the density matrix \hat{\rho} while the Lindblad (jump) operator \hat{L} describes the impact of the environment, which may cause a non-unitary time evolution. Choosing a spin- 1/2 system with \hat{H}=\Omega\hat{\sigma}_{y} , the unitary system dynamics itself (i.e., for \hat{L}=0 ) would correspond to the rotation between the states \left|\uparrow\right> and \left|\downarrow\right> as described above. Now let us add the Lindblad operator \hat{L}=\sqrt{\Gamma}\,\hat{\sigma}_{z} which generates decoherence with the damping rate \Gamma and leads to a decay of the coherences between the states \left|\uparrow\right> and \left|\downarrow\right> , i.e., the off-diagonal matrix elements of \hat{\rho} in the \left|\uparrow\right>,\left|\downarrow\right> basis (phase damping). Then the evolution equation for the expectation value of the spin operator \langle\mbox{\boldmath$\hat{\sigma}$}\rangle reads \displaystyle\frac{d}{dt}\langle\mbox{\boldmath$\hat{\sigma}$}\rangle=-2\mbox{\boldmath$\Omega$}\times\langle\mbox{\boldmath$\hat{\sigma}$}\rangle+2\mbox{\boldmath$L$}\times\left(\mbox{\boldmath$L$}\times\langle\mbox{\boldmath$\hat{\sigma}$}\rangle\right)\,, (2) with \mbox{\boldmath$\Omega$}=\Omega\mbox{\boldmath$e$}_{y} and \mbox{\boldmath$L$}=\sqrt{\Gamma}\,\mbox{\boldmath$e$}_{z} . For \Omega\gg\Gamma , we approximately find rotation with a frequency \Omega which slowly decays with a damping rate of \,{\cal O}(\Gamma) . For \Omega\ll\Gamma , on the other hand, the rotation disappears and we find a slow decay to the fully mixed state \hat{\rho}=\mbox{\boldmath$1$}/2 (with 1 denoting the unit matrix) on a long time scale of \,{\cal O}(\Gamma/\Omega^{2}) . As an intuitive picture, the environment measures the state of the system with an effective measurement rate \Gamma\sim 1/\Delta t . Thus, the slow decay to \hat{\rho}=\mbox{\boldmath$1$}/2 with the decay rate \,{\cal O}(\Omega^{2}/\Gamma) corresponds to the example above with the reduced transition rate \,{\cal O}(\Omega^{2}\Delta t) . Note that analogous results can be obtained for Lindblad operators in the form of projectors \hat{L}=\sqrt{\Gamma}\left|\uparrow\right>\left<\uparrow\right| or \hat{L}=\sqrt{\Gamma}\left|\downarrow\right>\left<\downarrow\right| or a combination of both \hat{L}_{1}=\sqrt{\Gamma_{1}}\left|\uparrow\right>\left<\uparrow\right| and \hat{L}_{2}=\sqrt{\Gamma_{2}}\left|\downarrow\right>\left<\downarrow\right| , see Appendix A . Furthermore, in App. B we also provide an example of Zeno freezing in the strong-coupling regime that does not require a Lindblad treatment. I.2 Adiabatic Quantum Algorithms In contrast to the usual scheme of sequential quantum algorithms nielsen2000 where the desired final state \left|\psi_{\rm out}\right> is obtained by applying a sequence of quantum gates onto some initial state \left|\psi_{\rm in}\right> , adiabatic quantum algorithms vandam2001a ; farhi2001a ; albash2018a encode the solution to the problem of interest into the ground state of a suitably chosen Hamiltonian \hat{H}_{\rm out} . In order to reach this final ground state \left|\psi_{\rm out}\right> , the idea is to start with an initial Hamiltonian \hat{H}_{\rm in} whose ground state \left|\psi_{\rm in}\right> is known and can be prepared efficiently and then to slowly deform the initial Hamiltonian \hat{H}_{\rm in} into the final \hat{H}_{\rm out} . A simple example is a linear interpolation scheme \displaystyle\hat{H}(t)=f(t)\hat{H}_{\rm in}+\left[1-f(t)\right]\hat{H}_{\rm out}\,, (3) with some function f(t) satisfying f(t=0)=1 and f(t=T)=0 where T is the run-time of the algorithm. If the change from \hat{H}_{\rm in} to \hat{H}_{\rm out} is slow enough, the adiabatic theorem ensures that we indeed end up in the desired final ground state \left|\psi_{\rm out}\right> . More precisely, an important condition for adiabatic evolution is that the matrix element of the rate of change \partial_{t}\hat{H} between two different instantaneous eigen-vectors \hat{H}(t)\left|\psi_{n}(t)\right>=E_{n}(t)\left|\psi_{n}(t)\right> with instantaneous eigen-energies E_{n}(t) is small sarandy2004a ; jansen2007a \displaystyle\left|\left<\psi_{n}\right|\partial_{t}\hat{H}\left|\psi_{m}\right>\right|\ll(E_{n}-E_{m})^{2}\,. (4) Assuming adiabatic evolution and starting in an eigen-vector \left|\psi(t=0)\right>=\left|\psi_{n}(t=0)\right> such as the ground state n=0 , the time evolved quantum state \left|\psi(t)\right> stays close to the instantaneous eigen-vector \left|\psi_{n}(t)\right> up to phase factors containing the dynamical phase \dot{\varphi}_{n}^{\rm dyn}(t)=E_{n}(t) and the geometrical phase \dot{\varphi}_{n}^{\rm geo}(t)=i\left<\psi_{n}(t)\right|\partial_{t}\left|\psi_{n}(t)\right> . I.3 Adiabatic Quantum Search Algorithm As a simple example for such an adiabatic quantum algorithm, let us consider for a single solution the adiabatic version of Grover’s quantum search algorithm grover1997a as proposed by Roland and Cerf roland2002a . In the Grover search, the task is to find the solution state \left|w\right> as one of the N computational basis states (such as \left|w\right>=\left|01\dots 10\right> ) within an unsorted database. Since we obviously do not know the solution \left|w\right> beforehand, we choose the superposition state \left|s\right>=\sum_{x=0}^{N-1}\left|x\right>/\sqrt{N} as the initial state. In an effective spin representation with n spins and N=2^{n} basis states and identifying \left|\uparrow\right>=\left|1\right> and \left|\downarrow\right>=\left|0\right> , this state \left|s\right>=[(\left|0\right>+\left|1\right>)/\sqrt{2}]^{\otimes n} has all spins aligned in