Although Zadeh may have had the idea "of the best fit POSSIBLE" in the back
of his mind, that is hardly how I would characterise the motivation for
using fuzzy logic.For me it brings science closer to the way that people,
including scientists, think, and thus can do a better job of approximating
the best possible reasoning.
Common examples can be found in some of the fuzzy decision tools that aare
widely circulated as demos. For example, if you are buying a house and your
rules include:
IF the house is close to work AND not too expensive THEN ...
the usual approach would be to define "close to work" and "too expensive" as
perhaps <10 km and >200.000 €, but a real person would probably settle for a
really great buy 12 km from work or a 220.000 € house within walking
distance. This is easily represented by treating these concepts with fuzzy
memberships.
Another reason for using fuzzy logic is that it is simple. Literally so. The
first applications of fuzzy logic were in system control, because back in
the days before cheap high-speed processors it was possible to build
economical controllers to replace slow and expensive systems based on
traditional methods with systems of differential equations to be solved.
Unfortunately there are those who feel that any mathematical approach has to
be fancy and sophisticated and involve lots of equations, which is one of
the reasons that fuzzy logic has been slow to be adopted in ecology, but it
is really very simple.
Bill Silvert
----- Original Message -----
From: "Wayne Tyson" <landr...@cox.net>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, August 06, 2009 9:22 PM
Subject: Re: [ECOLOG-L] Fuzzy Logic in Ecology
Re: Ecology Logic Fuzzy*
I was never much good at statistics, to put it mildly (two or three courses
put a real drag on my GPA and my intuition).
It seems (intuitively) to me that Bayes and Zadaeh both had good ideas that
were aimed at the direction of the best fit POSSIBLE, the former seeming to
be more elegantly conforming to hypotheses, the latter embracing uncertainty
more certainly--maybe.
I am grateful for those laboring in the fields of numerical research;
without them observers of the interplay of organisms and their environments
might tumble down some warren of fantasy and never to return. Yet, looking
in the looking-glass never hurt anybody.
On the other hand, it seems that the Sword of Certainty, especially in the
form of irrelevant decimal points and other attempts to reduce complex and
constantly-changing phenomena to quantities that fit neat boxes carved out
for them, hangs threateningly over jiggly and evasive phenomena, threatening
to freeze it all in place for all time. Still, provided analyses don't stray
too far from the actual data manipulated and are not treated as license for
(ironically?) unbridled conjecture amongst self-ordained emperors, those
numbers can open up insights as well as interfere with them. I have no firm
answer for this conundrum.
What are the Great Questions in ecology? What are the answers? What answers
have actually changed rather than just been re-clothed or bumped off the
mountain in, say, the last century or so? What new questions have been
added?
WT
*Sorry, Bill, in a lame attempt to please everybody I didn't fuss with the
subject line (what a wuss, eh?), I was going to ask Eric off-list but the
reply-to-all button didn't have his email address in it, so I just deleted
one of the apparently duplicate list addresses--gad, this is getting
complicated! Back to square one . . .
----- Original Message -----
From: "William Silvert" <cien...@silvert.org>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, August 06, 2009 8:13 AM
Subject: [ECOLOG-L] Fuzzy Logic in Ecology
I missed Wayne's posting since the subject line had nothing to do with fuzzy
logic, and although I know that I will raise a chorus of outrage I am
changing the subject line.
Quite a few papers on fuzzy logic aplications have been published, many in
Ecological Modelling (including a couple of my own). I also have some
PowerPoint presentations on my website, http://ciencia.silvert.org, and
there are many in related fields such as soil science. I think that one of
my first efforts was in niche theory, obviously the "multi-dimensional
manifold" that Hutchinson defined must have fuzzy boundaries.
Unfortunately a lot of work in the field seems to consist of lots of
mathematics with little biological content. I like to think in terms of
fuzzy rules like
IF the weather is warm AND nutrient levels are high THEN there is a serious
risk of bottom anoxia
(which is similar to a rule composed for aquaculture siting) but many
authors seem more concerned with whether the membership function is
triangular or trapezoidal.
As with any new field or approach, fuzzy ecology is experiencing growing
pains. There was an international conference on the topic in Kiel many years
ago, perhaps 15, but progress has been slow.
Bill Silvert
----- Original Message -----
From: "Jonathan Nelson" <eco...@phossie.com>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, August 06, 2009 3:00 PM
Subject: Re: [ECOLOG-L] Bayesian analysis in population ecology workshop,
early registration deadline 10 August 2009
On Wed, Aug 5, 2009 at 7:25 PM, Wayne Tyson <landr...@cox.net> wrote:
Do you know if anyone has investigated the possible application of fuzzy
logic theory to ecology/population ecology and why or why not?
Re the first part of the question:
Results 1 - 10 of about 12,300 for fuzzy logic ecology.
http://scholar.google.com/scholar?q=fuzzy%20logic%20ecology&oe=utf-8
The results include the 2002 overview paper by Regan, Colyvan, Burgman
(Ecological Applications, 12(2), 2002, pp. 618–628) as the 3rd result,
available as PDF. I had not seen this before but I'll be reading it
this evening, as it looks fascinating:
"Abstract. Uncertainty is pervasive in ecology where the difficulties
of dealing with
sources of uncertainty are exacerbated by variation in the system
itself. Attempts at clas-
sifying uncertainty in ecology have, for the most part, focused
exclusively on epistemic
uncertainty. In this paper we classify uncertainty into two main
categories: epistemic un-
certainty (uncertainty in determinate facts) and linguistic
uncertainty (uncertainty in lan-
guage). We provide a classification of sources of uncertainty under the
two main categories
and demonstrate how each impacts on applications in ecology and
conservation biology.
In particular, we demonstrate the importance of recognizing the effect
of linguistic uncer-
tainty, in addition to epistemic uncertainty, in ecological
applications. The significance to
ecology and conservation biology of developing a clear understanding
of the various types
of uncertainty, how they arise and how they might best be dealt with
is highlighted. Finally,
we discuss the various general strategies for dealing with each type
of uncertainty and offer
suggestions for treating compounding uncertainty from a range of sources."
--------------------------------------------------------------------------------
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.392 / Virus Database: 270.13.44/2283 - Release Date: 08/05/09
05:57:00
