home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky alt.control-theory:16 comp.ai.neural-nets:4750 comp.realtime:1522 comp.theory.dynamic-sys:401 comp.theory.self-org-sys:135
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!usc!randvax!hermix!news.cerf.net!nic.cerf.net!hlf
- From: hlf@nic.cerf.net (Howard Ferguson)
- Newsgroups: alt.control-theory,comp.ai.neural-nets,comp.realtime,comp.theory.dynamic-sys,comp.theory.self-org-sys
- Subject: Re: Use of hysteresis in control problems
- Keywords: artificial neural networks,control
- Message-ID: <1iilfcINN6q0@news.cerf.net>
- Date: 8 Jan 93 01:30:20 GMT
- References: <1993Jan2.191649.874@noose.ecn.purdue.edu> <1993Jan5.200214.4270@news.acns.nwu.edu>
- Sender: hlf@cerf.net
- Distribution: inet
- Organization: CERFnet Dial n' CERF Customer Group
- Lines: 33
- NNTP-Posting-Host: nic.cerf.net
-
- In article <1993Jan5.200214.4270@news.acns.nwu.edu> g-yuen@nwu.edu writes:
- >
- >Some of my recent work suggest hysteresis may be a fairly common
- >features of real neurons. My colleagues often mentioned that this
- >is a useful property for some control problems, but has never
- >presented any convincing examples or arguments for its use. Can
- >someone out there in netland give me some concrete examples or
- >references to get into ?
- >
- >Your response will be greatly appreciated.
- >
- >
- >
-
- Well I know of one good example, but it is not quite a real time
- solution, but more of a GUI situation. In the project I am currently
- working on we display a graphic on the screen which represents
- a length of time. The operator can expand this length of time or
- shrink it gradually. ( All of the smaller components of the time
- period can be changed also but that is not relavent here).
-
- Because the graphic gets too cramped when shrunk it must be rescaled
- ocationally. Now if the operator is trying to adjust the length to be
- almost equal to the rescale point then he may get a flicker effect
- where the graphic switches between the larger and smaller graphic
- several times as he is trying to make this fine adjustment. The
- solution is to make the big-to-small rescale happen at a value
- which is larger than the small-to-big rescale point.
-
- It is kind of hard to explain all that without pictures, but I hope
- you get the idea,
-
- hlf
-