By 'wrong reference state' I mean a state which cannot be transformed into desired ones via variational ansatz

$\left|\Psi\left[\mathbf{n}\right]\right\rangle =e^{i\hat{O}\left[\mathbf{n}\right]}\left|ref\right\rangle \;,\; H\left[\mathbf{n}\right]=\left\langle \Psi\left[\mathbf{n}\right]\right|\hat{H}\left|\Psi\left[\mathbf{n}\right]\right\rangle $

where $\mathbf{n}$ is the variational parameter (field).

To be concrete, if we started with the first excitation state in a gapped system, i.e. choosing (making a guess) this first excitation state to be the reference state, then it seems to me that we will never reach the true ground state, as well as the correct excitations. The manifolds are disconnected because of the gap.

Are there any concrete examples for my general question here? Maybe I have not properly described this question. Please let me know if you are interested but don't understand. I will try to clarify.


If the true ground state is not in (or close to) the manifold of states over which the variation is done, the solution found will simply be meaningless.

Thus a good numerical scheme will have an accompanying analysis that the approximation scheme is at least in principle able to approximate the desired solution.

  • $\begingroup$ Actually I am working on this article - "Spontaneous interlayer coherence in double-layer quantum Hall systems: Charged vortices and Kosterlitz-Thouless phase transitions" Phys. Rev. B 51, 5138–5170 (1995) prb.aps.org/abstract/PRB/v51/i8/p5138_1 I am not sure about the ferromagnetic ground state they used. They may have strong supports from both numerical and experimental observations. But currently people also check other possibilities, for example the "entanglement spectrum" study of QH systems. So the philosophy behind their guess seems quite crucial for their entire argument. $\endgroup$ – Yunlong Lian Nov 19 '12 at 13:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.