Differences
This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revision | |
| documentation:language_reference:objects:responsefunction:functions:reducepoles [2025/11/20 04:20] – external edit 127.0.0.1 | documentation:language_reference:objects:responsefunction:functions:reducepoles [2026/11/13 07:51] (current) – Maurits W. Haverkort |
|---|
| ### | ### |
| //Gout = ResponseFunction.ReducePoles(Gin, NPoles)// or //Gout = ResponseFunction.ReducePoles(Gin, NPolesVal, NPolesCon)// reduces the number of poles in $G_{in}$ to $N_{Poles}$ or to $N_{Poles}^{Val} + N_{Poles}^{Con}$ by merging poles close to each other with small spectral weight (residue) (see appendix of [[https://doi.org/10.1103/PhysRevB.90.085102||Phys. Rev. B 90, 085102]] for additional information). The algorithm with $N_{Poles}^{Val}$ and $N_{Poles}^{Con}$ set does not shift spectral weight between the conduction and valence states (set by the chemical potential). | //Gout = ResponseFunction.ReducePoles(Gin, NPoles)// or //Gout = ResponseFunction.ReducePoles(Gin, NPolesVal, NPolesCon)// reduces the number of poles in $G_{in}$ to $N_{Poles}$ or to $N_{Poles}^{Val} + N_{Poles}^{Con}$ by merging poles close to each other with small spectral weight (residue) (see appendix of [[https://doi.org/10.1103/PhysRevB.90.085102||Phys. Rev. B 90, 085102]] for additional information). The algorithm with $N_{Poles}^{Val}$ and $N_{Poles}^{Con}$ set does not shift spectral weight between the conduction and valence states (set by the chemical potential). |
| | |
| | A response function in a representation with a chain (//Tri//, //Nat// or //DoubleTri//) can be made smaller in a different way, by cutting the chain with the option //NTriMax// of //[[documentation:language_reference:objects:responsefunction:functions:changetype|ResponseFunction.ChangeType()]]//. A chain cut to $N$ steps keeps the first $2N$ moments of the response function. |
| ### | ### |
| |