Several questions have arisen concerning the interior of the spectrum of an operator. The answers and solutions to these questions are very similar or related. I feel we should wikify these answers.

Examples are:

How to find the interior eigenvalues by krylov subspace method?

What is the fastest way to calculate the largest eigenvalue of a general matrix?

Using algebraically smallest eigenvalues to find smallest in magnitude eigenvalues

Fast algorithms to find the eigenvalues of some matrix on intervals of interest


  • 1
    I guess I'd have to look at the questions. Do you have a list of the questions you feel should be wikified? In general, I'm against wikifying questions, aside from maybe the occasional "big list of resources" question where lots of people contribute. Wikifying answers is another story. – Geoff Oxberry Apr 11 '12 at 16:54
  • @GeoffOxberry: I added a few links. – Deathbreath Apr 11 '12 at 20:24
  • These questions are definitely not duplicates. The topic is huge, and small changes to the problem setup open up different solution methods. One could fill a small library with literature about calculating eigenvalues. – Nick Alger Jul 20 '14 at 17:16

I don't think wikification is an appropriate action here. Basically, community wiki is used for answer posts that get edited a lot, get edited by multiple users, or denote that a post has substantial contributions from multiple users. For a discussion of what community wiki means, see "What are 'Community Wiki' posts?" on Meta Stack Overflow, and Grace Note's post on "The Future of Community Wiki". The issue you're raising is not collaborating, but similarity, possibly verging on duplication.

I'd definitely leave the second and third questions on your list alone, because they don't duplicate anything, to my knowledge.

There is an argument to be made that the first and fourth questions on the list are duplicates. I hesitate to flag them as duplicates because there are eigenvalue algorithms out there that are not Krylov subspace methods. If someone can make a persuasive argument with some references that says that the only fast algorithms to find eigenvalues of a matrix in an interval of interest are Krylov subspace methods, then I'll mark the fourth question as a duplicate of the first (or vice versa). However, lacking a broad understanding of the different types of eigenvalue methods out there, I'd rather leave them alone until someone makes that argument.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .