Topological Discrimination of Steep to Supersteep Gap to Emergence of Tunneling in Adiabatic Quantum Processes

arxiv(2023)

引用 0|浏览2
暂无评分
摘要
It is shown that the gap that limits the speed of a quantum annealing process can take three salient morphologies: (i) the supersteep gap where both the ground and the first excited eigenenergy level curves have topologically related pairs of nearby inflection points giving both the maximum and the minimum a steep aspect, (ii) the steep gap where only the first excited eigenenergy level curve has a pair of inflection points giving its minimum a steep aspect while the maximum of the ground level does not exhibit inflection points, and (iii) the mild gap that has no related inflection points. Classification of the various singularities betrayed by the inflection points relies on the critical value curves of the quadratic numerical range mapping of the matrix H0+iH1, where H0 is the transverse field Hamiltonian and H1 the problem Hamiltonian. It is shown that the ground level is mapped to the generically smooth boundary of the numerical range, while the first excited level is mapped to an interior non-smooth critical value curve exhibiting swallow tails. The major result is that the position of the swallow tails relative to the boundary allows the supersteep versus steep discrimination, while the absence of swallow tail-boundary interaction characterizes the mild gap. As a corollary of the singularity analysis, the highly structured initial and final Hamiltonians of the Grover search create unstable singularities that break into stable swallow tails under perturbation, with the consequence of invalidating the gap scaling estimates computed around the unstable singularity. Classification of all stable singularities from a global viewpoint requires the Legendrian approach where the energy level curves become Legendrian knots in the contact space. Last but not least, it will be shown that a supersteep swallow tail previews tunneling.
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要