Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!wuarchive!m.cs.uiuc.edu!ibma0.cs.uiuc.edu!sunc6.cs.uiuc.edu!noe
From: noe@sunc6.cs.uiuc.edu (Roger Noe)
Newsgroups: comp.theory
Subject: Re: Closed form solution to a recurrence
Message-ID: <283CA186.1F92@ibma0.cs.uiuc.edu>
Date: 24 May 91 05:15:50 GMT
References: <9105231143.AA25697@dali.uni-paderborn.de> <1991May24.105457.1@csc.anu.edu.au>
Sender: news@ibma0.cs.uiuc.edu
Organization: University of Illinois at Urbana-Champaign
Lines: 11

In article <1991May24.105457.1@csc.anu.edu.au> bdm659@csc.anu.edu.au writes:
>If C(n,k) denotes the binomial coefficients, taken to be 0 except
>where 0 <= k <= n, then   A(i,k) = C(i-k,k-1) + C(i-k,k).

Which is of course just C(i-k+1,k).  This assumes that A(i,0)=1, the
unspecified boundary condition not covered by the recurrence relation.
--
Roger Noe                            roger-noe@uiuc.edu
Department of Computer Science       noe@cs.uiuc.edu
University of Illinois               40:06:39 N.  88:13:41 W.
Urbana, IL  61801  USA