Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site allegra.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!ask
From: ask@allegra.UUCP (A.S. Krishnakumar)
Newsgroups: net.math
Subject: Re: Some math questions
Message-ID: <2878@allegra.UUCP>
Date: Tue, 30-Oct-84 11:28:28 EST
Article-I.D.: allegra.2878
Posted: Tue Oct 30 11:28:28 1984
Date-Received: Wed, 31-Oct-84 00:30:33 EST
Distribution: net.math
Organization: AT&T Bell Laboratories, Murray Hill
Lines: 10


	A correction to the note of ...browngr!jfh regarding square root
of a matrix:  A real symmetric matrix has real, but not necessarily
non-negative eigenvalues.   The square root algorithm given by jfh
requires that A be positive semi-definite.  It is easy to see why it must
be so.  The eigenvalues of A are the squares of those of B and hence must
be non-negative (this assumes, of course, that we are restricting ourselves
to real matrices).  With this proviso, the algorithm is correct.

							Krishna