[From Rick Marken (920304)]
Bruce Nevin (Wed 92014 10:35:52) says:
To point this up, the box on p. 38 of the article describes Takeshi
Yamakawa's fuzzy controller and its solution of the "inverted pendulum
control problem."
Can we show how HPCT maintains the virtual equivalents of fuzzy sets and
does the equivalent of fuzzy inference for this sort of control?
I think the best comment on this fuzzy control stuff was made about a
year ago when the fuzzy "inverted pendulum controller" was posted to
csgnet. This program was huge (in source) -- many megabytes. Bill
Powers (in a matter of hours) posted an inverted pendulum controller
based on good old fashioned control theory. It was maybe twenty lines
of code, most of which handled the physics. My opinion -- fuzzy logic
is a solution looking for a problem. The problems that it has been
applied to already have better solutions. But fuzzy logic is trendy
science -- so that's where the bucks go.
I think you are right -- people seem to want to deal with the world at
the principle level or higher. I see this here at Aerospace where there
is an unbelievable reverence for "expert systems" -- and this is
coming from engineers, many of whom have been trained in control theory.
I think they are so used to dealing with control problems verbally -- like
"That voltage there is decreasing TOO QUICKLY. We better switch to SIDE B"--
that they don't notice that they are just controlling quantitative variables
(keep the rate of dischage = x). It's a very interesting phenomenon, this
regression to symbol manipulation by quantitatively trained people. I
think it speaks to how deeply imbued we all are with stimulus-response
preconceptions about behavior.
To Bill Powers: You may have done this before but I would appreciate it
if you could post a simple diagram of the "persons" in the Gatherings
program. I want to write a program that demonstrates (on an individual
basis) some of the points you make with the gatherings program -- i.e.
behaviors like "circles of people" and "backtracking" are just observed
side effects of each person's efforts to control peceptions like
proximity, speed, etc. So what are the controlled variables in the
Gatherings program and how are they related? I think I remember
that one was "proximity" and the other is "speed" -- is that right?
Are they controlled independently? Thanks for the info.
Regards
Rick
ยทยทยท
**************************************************************
Richard S. Marken USMail: 10459 Holman Ave
The Aerospace Corporation Los Angeles, CA 90024
Internet:marken@aerospace.aero.org
(310) 336-6214 (day)
(310) 474-0313 (evening)