home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!sunic!dkuug!diku!torbenm
- From: torbenm@diku.dk (Torben AEgidius Mogensen)
- Newsgroups: comp.theory.cell-automata
- Subject: Re: cellular automata with different rules at each site
- Message-ID: <1993Jan22.120439.24866@odin.diku.dk>
- Date: 22 Jan 93 12:04:39 GMT
- References: <C188rz.2Jq0@austin.ibm.com>
- Sender: torbenm@thor.diku.dk
- Organization: Department of Computer Science, U of Copenhagen
- Lines: 30
-
- demaris@austin.ibm.com (Dave Demaris) writes:
-
-
- >These are called inhomogeneous cellular automata.
-
- I would expect that any automaton with a finite set of different rules
- for different sites can be simulated by a uniform automata with a
- somewhat larger state-space.
-
- A simpe example is wire-world. It can be seen as having different
- rules for background cells (one possible state) and wire cells (three
- possible states), but usually it is implemented as a single type of
- cell having four possible states.
-
- It can, though, be easier to describe an automaton as having several
- types of cells with each their state space rather than a single type
- of cell with a huge state-space.
-
- A truly inhomogenous automaton (which can't easily be simulated with a
- uniform automaton) is one where each cell is not connected to it's
- nearest neighbours but to arbitrary other points. If there is no limit
- to the distance, there is no obvious way to reduce this to a uniform
- grid. In a similar vein one could imagine an automaton on a
- non-periodic Penrose tiling. Using the variant with two rhombic tiles,
- each will have four neighbours, so the rules for each cell could be
- identical, but the global behaviour would be non-uniform.
-
- Torben Mogensen (torbenm@diku.dk)
-
-
-