home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!sgiblab!sgigate!rutgers!igor.rutgers.edu!remus.rutgers.edu!clong
- From: clong@remus.rutgers.edu (Chris Long)
- Newsgroups: rec.puzzles
- Subject: Re: Gale and Bridges (PARTIAL SPOILER)
- Message-ID: <Nov.22.01.54.49.1992.14102@remus.rutgers.edu>
- Date: 22 Nov 92 06:54:49 GMT
- References: <1992Nov20.191537.27029@cc.ic.ac.uk>
- Organization: Rutgers Univ., New Brunswick, N.J.
- Lines: 19
-
- In article <1992Nov20.191537.27029@cc.ic.ac.uk>, Nairo Aparicio writes:
-
- > Here there is a problem published on one of the boards of this department:
- > "In a certain wide river, there are 6 islands joined to each other and
- > the banks by 13 bridges, as shown below. Unfortunately if there is a
- > gale, the bridges will each collapse with independent probability 1/2.
- > What is the probability that the river will be crossable after the gale?"
-
- The river will be crossable iff there is a path from top to bottom;
- it will be uncrossable iff the river has a path from left to right.
- In the case for a rectangular bridge grid of mxn islands with probability
- of collapse p the associated river grid is also rectangular but with size
- (n-1)x(m+1) and (essentially) probability of collapse of 1-p. Therefore,
- P(m,n,p) = 1-P(n-1,m+1,1-p); for the given problem we get P(2,3,1/2) =
- 1-P(2,3,1/2) ==> P(2,3,1/2) = 1/2. For general p the problem is more
- tedious; I haven't done the computation yet, but I get that it's a
- polynomial of degree 13 in p.
- --
- Chris Long, 265 Old York Rd., Bridgewater, NJ 08807-2618
-