Answers are grounded in this paper and cite their sources.
A new technique is developed in this paper for solving Hamiltonians of the form $H={Σ}{}{k>0}{{ε}}ₖ({{a}ₖ}⁺{a}ₖ+{{a}_{{-}k}}⁺{a}_{{-}k}){-}{Σ}{}{k,l>0}G(k,l){{a}ₖ}⁺{{a}_{{-}k}}⁺{a}_{{-}l}{a}ₗ,$ where $G(k, l)$ may have the forms: $G(k, l)=G$ (independent of $k$ and $l$), $G(k, l)={G}ₖ{G}ₗ$, or $G(k, l)=G(k, l)$ (nonseparable in $k$ and $l$). The eigenvalues obtained with this technique are compared with the exact solutions of Kerman, Lawson, and Macfarlane. The agreement with the exact solutions is fairly good.
No takes yet. Share an insight, caveat, or question.
R.R. Chasman (1963) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: