Newsgroups: comp.theory,comp.graphics.algorithms,comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!canopus.hbs.edu!news-in.tiac.net!uunet!in3.uu.net!news2.new-york.net!abigail
From: abigail@ny.fnx.com (Abigail)
Subject: Re: Dijkstra algo on planar (grid) graphs
X-Perl-Wrapper: Version 0.1 (by Abigail)
X-Newsreader: slrn (0.9.0.0 (BETA) UNIX)
X-Nntp-Posting-User: abigail
Reply-To: abigail@ny.fnx.com
X-Http: http://www.ny.fnx.com/abigail/
Lines: 18
Organization: FNX Ltd, Intelligent Risk Management
Message-ID: <E0Dq8K.Mqu@news2.new-york.net>
References: <AB62WVo00H@creatgraphics.spb.su>
X-Trace: 847166993/29353
X-Nntp-Posting-Host: melgor.ny.fnx.com
Date: Tue, 5 Nov 1996 04:09:55 GMT
Xref: glinda.oz.cs.cmu.edu comp.theory:17226 comp.graphics.algorithms:39992 comp.ai.games:6845

On 4 Nov 1996 18:00:25 +0300, Andrew Iones wrote in comp.theory,comp.graphics.algorithms,comp.ai.games:
++ Hello,
++ 
++ Have you heard anything about a modification of Dijkstra algorithm
++ for planar or grid graphs? It seems that there exists an algo
++ with lower complexity then O(n*n) (without any tricky data storage
++ resulting in O(nlog(n)) complexity).


Not sure what you mean by n. Number of vertices? Anyway, Dijkstra's
algorithm runs in O (|V| lg |V| + |E|), and for planar graphs,
|E| = O (|V|). So, without any modifications, Dijkstra's algorithm runs
in O (|V| lg |V|).




Abigail

