Reply to "Comment on `Generalized James' effective Hamiltonian method'"

In the preceding Comment [1] it was claimed that the third-order Hamiltonian obtained in our original paper [2] is not Hermitian for general situations when considering time-dependence and the way of deriving the effective third-order expansion is not very rigorous. To reply the comment we should emphasize the following three points: first of all, the third-order Hamiltonian given in our paper is exactly Hermitian under the conditions mentioned there. Secondly, the iterative method adopted in our paper to derive the generalized effective Hamiltonian is equivalent to the Dyson series, and its correctness can thus be guaranteed. Thirdly, although the truncated effective Hamiltonian is indeed non-Hermitian under the time-dependent situation as presented in the Comment, it corresponds exactly to the non-unitary truncated Dyson series. Considering the truncated Dyson series has been extensively utilized in the time-dependent perturbation theory, in our opinion, the non-Hermitian truncated effective Hamiltonian can still be treated as an approximation of the effective Hamiltonian.