Systems of systems - Ken Udut wonders - what is the terminology for it?
Ken Udut
Ken Udut
There must be a word or phrase for a "system of systems" - that is,
the science or perhaps classification of various systems of the world.
Udut Ken
There are systems of philosophy, systems of science, systems of
religion, systems of logic, systems of mathematics - pick something,
and there are systems.
Could it be the study of metasystems? I don't know. That would
simply be another system, no?
Can anybody shed a little light on this for me? It would be much
appreciated!
Ken
Doug Elias, Ph.D.
In article <b62c7b9b.01121...@posting.google.com>, simplify3
@aol.com says...
The "science" of systems is "Systems Science", which has as its focus of
interest the study of the characteristics of the relationships among
systemic components, deliberately excluding characteristics of the
components themselves (this latter is deemed to be the focus of the
"specific disciplines", e.g., "chemistry", "biology", etc.).
There are a fair number of different approaches to the problem of
"classification of systems" ... I would recommend a good introductory
text on Systems Science as a starting point, one that gives an overview
of a number of different schools of thought; I believe that Dr. George
Klir wrote one not too long ago ... yes, "Facets of Systems Science",
you can find it listed at Amazon or BooksAMillion.
However, one word of caution: there are *no* "systems" without pre-
existing *problems* ... i.e., a system gets defined in relation to a
focus of attention, the need to address some issue. This is one of the
reasons why there are different schools of Systems Thinking: each
focuses on different types of issues, or different views of the issues,
or different methodologies using which the issues might be addressed
(these latter schools could be called "methodology-driven").
Where it *really* gets fun is in working towards a classification of
these classifications ... which becomes significant if you're trying to
find a "good" set of systems-tools for a particular systems-problem.
This is somewhat analogous to which toolbox an all-purpose handyman
takes on a particular job: taking the plumbing toolbox on a roofing job
is a non-starter, and it's very much the same in the field of General
Systems Problem Solving -- you have to fit the methodological tools to
the constraints of the problem/system.
Please let me know if there's anything more I can tell you ... I've been
in this field for quite some time, as student, practitioner and
professor, and it still gives me goosebumps.
doug
Ken Udut
> There are a fair number of different approaches to the problem of
> "classification of systems" ... I would recommend a good introductory
> text on Systems Science as a starting point, one that gives an overview
> of a number of different schools of thought; I believe that Dr. George
> Klir wrote one not too long ago ... yes, "Facets of Systems Science",
> you can find it listed at Amazon or BooksAMillion.
I do believe that I'll be getting myself an early christmas present!
[Booksamillion is one of my favorite book ordering sites, too]
> However, one word of caution: there are *no* "systems" without pre-
> existing *problems* ... i.e., a system gets defined in relation to a
> focus of attention, the need to address some issue. This is one of the
> reasons why there are different schools of Systems Thinking: each
> focuses on different types of issues, or different views of the issues,
> or different methodologies using which the issues might be addressed
> (these latter schools could be called "methodology-driven").
True. I had never thought of the reason behing systems is an attempt
to solve a problem. Then there are those linkages inbetween such
systems, which sometimes don't come out until after a more specific
system has been developed. Oh wait - you address that below... :-)
> Where it *really* gets fun is in working towards a classification of
> these classifications ... which becomes significant if you're trying to
> find a "good" set of systems-tools for a particular systems-problem.
Yes - systems of systems! That's what I'm searching for. If not a
skeleton-key that fits many locks, at the very least, a system that
answers many questions - provides tools for many 'pipes'.
I suppose some of the cross-pollinization from area-to-area of human
knowledge is a way that this is occuring in small ways.
> Please let me know if there's anything more I can tell you ... I've been
> in this field for quite some time, as student, practitioner and
> professor, and it still gives me goosebumps.
>
> doug
It's my cup of tea as well, although I am just a corporate Excel guru,
blessed enough to have created a little niche doing work that's just
challenging enough, but leaving me enough brain-room to ponder.
-Ken
Doug Elias, Ph.D.
[">>" is me, ">" is Ken]
> > Where it *really* gets fun is in working towards a classification of
> > these classifications ... which becomes significant if you're trying to
> > find a "good" set of systems-tools for a particular systems-problem.
>
> Yes - systems of systems! That's what I'm searching for. If not a
> skeleton-key that fits many locks, at the very least, a system that
> answers many questions - provides tools for many 'pipes'.
This is strictly terminology, so that we'll understand what is meant by
what we say: a "system of systems" is, to me, a higher-order system that
is composed of lower-order systems, in the sense of:
...
* a super-super-system is composed of super-systems,
* ...each of which is composed of systems
* ...each of which is composed of sub-systems,
* ...each of which is composed of sub-sub-systems,
...
So that what you've got is potentially infinite regress both above and
below, and a given entity can, at one and the same time, be considered a
"system" due to its being the central focus of an investigation, a
super-system because one of its component systems is the main focus, a
sub-system since a (larger-scale) system it is a part of is the main
focus, etc. etc.
What I'm trying to describe by "classification of classifications" is,
rather, "a system of system-classification-schemes" ... I'm looking for
what each Way, each different Systems Theory, contributes to the overall
understanding regarding the entity being investigated, and where these
knowledge-contributions overlap (meaning that two or more Systems
Theories share a particular viewpoint), and where they are unique. Once
this is in place, I can take a particular problem and say "I need to
have this kind of knowledge about this problem", go to my classification
structure, see what Systems Theories incorporate that kind of knowledge,
apply one or more of them to the problem (in the sense of using the
methodological tools that exemplify them), and fit the resulting systems
knowledge into the resolution I'm constructing for the problem.
> I suppose some of the cross-pollinization from area-to-area of human
> knowledge is a way that this is occuring in small ways.
This is one of the major strong points of Systems Science: the ability
to recognize that a set of relationships within one type of system, the
ways that the components of that system interact with one another, are
*isomorphic* to the set of relationships within a completely different
type of system ... so, even though the actual components of the two
systems are totally different, you can treat them as identical to one
another because they interact with their peer-components in exactly the
same way. A "circuit" will function the same way regardless of whether
it is built out of mechanical components, electro-magnetic components,
or fluidic components, as long as all of those different kinds of
components are related to the rest of their respective system-components
according to the proper physical laws -- because the systems of laws
that govern their behavior are *identical* in form. This allows you to
transfer knowledge gained in one form of system or context, to another
form ("cross-pollination", as you call it) without having to do all of
the same groundwork in the latter that you've already done in the former
... you simply make sure that you relate the system-components
correctly.
I'm in the process of writing up class-notes for a graduate-level course
I'm teaching next semester on one approach to this problem, one
particular Systems Theory. I'll post excerpts here, and maybe we can
get some discussions going on these and related topics.
doug
George R. McConnell
That was a good reply to your problem I would certainly endorse the
Klri reference - just one thing to add. The term "System of Systems"
is coming to have a quite specific meaning. This is primarily from
the defence world and I have to say that I am not entirely in
agreement with its use, however it is as well to be aware that the
term has a quite specific meaning for some people.
regards
George
Doug Elias, Ph.D.
In article <qrf72us5lhr6ngs4l...@4ax.com>,
geo...@nospam.frenchhorn.screaming.net says...
> The term "System of Systems"
> is coming to have a quite specific meaning. This is primarily from
> the defence world and I have to say that I am not entirely in
> agreement with its use, however it is as well to be aware that the
> term has a quite specific meaning for some people.
I've already mentioned to Ken in an earlier reply that, to me, "system
of systems" is basically a shorthand way of describing something that is
made up of things that can be considered "systems" in their own right,
so "system of systems" would be analogous to "sets of subsets", ad
infinitum both above ("super-supersets of supersets of sets of ..." and
below ("...subsets of sub-subsets of ..."). The central point is to
emphasize that any given entity can, at one and the same time, be
considered to be:
* comprised, in part, of the thing you're really interested in right now
(thus a "super-system" of the "system-of-interest"),
* the thing you're really interested in right now (thus the "system-of-
interest"), and
* a component of the thing you're really interested in right now (thus a
"sub-system" of the "system-of-interest"),
...and it's all in the eye of the investigator, 'cause the entity itself
never changes -- the focus of the investigator's attention is the
determinant of where the entity rests in the systemic hierarchy.
I would be interested to hear how else this phrase is used, and the
rationale.
doug
ps...Happy New Year to all, and may those with violence in their hearts
be shown the damage they do to their souls.
Robert Haviland
has no system properties, and is represented by such ideal as L, C, etc.
The number of levels is still beinmg worked out. Seven may be enough, but at
one time the Air Force (Wright field) called for 12, some of which were
management levels, which seem to be a separate field.
R. P. Haviland
qualsyste...@mindspring.com
One possible "teminology" that deals with systems-of-systems is that
described by Complexity Theory. This is a body of knowledge that has
yet to condense into a full, truly unified theory but much work has
already been done at the Sante Fe Institute and others.
Complex Systems (or Complex Adaptive Systems) can be comprised of
systems of systems and quite frequently one wants to control or at
least encourage the output(s) produced by these systems in some way.
The exciting thing for me is that the field of Complexity is intended
to provide tools that deal with some of the most intractable
problems/systems known to man.
An interesting concept that is recognized by complexity practioners is
that of emergence. This describes the observation that some systems
contain properties at the system level that do not exist at the part
level. For example, although individual human brain cells do not
contain the property of consiousness, if enough of them are brought
together and organized into a brain you now have the property of
consciouness where before you did not. Another example is that of
automobile traffic patterns. Grid-lock does not apply to the idvidual
parts that make up a traffic system (the roadways, drivers, wheels,
gasoline, pistons, rainfall etc.) but when these and other parts are
allowed to form a "system", grid-lock can certainly occurr.
One may describe a spectrum of systems as deterministic (those that
can easily be described by relatively simple equations), complicated
(those made up of several systems and perhaps redundancy) and chaotic
systems (those that behave randomly). Complex systems live somewhere
between complicated systems and chaotic systems.
I by no means am an expert on this topic but can see the value and
need for more methods that can deal with multi-layered systems
containing parts that can interact with other parts as well as its
environment.
Steve
George R. McConnell
<dm...@cornell.edu> wrote:
>I would be interested to hear how else this phrase is used, and the
>rationale.
>
Hi Doug,
Here are a few links to web sites which cover the 'alternative' view
of system of systems.
http://www.isr.umd.edu/~selberg/SystemOfSystems/
http://www.aero.org/capabilities/systemsEngineering.html
http://www.rand.org/publications/IP/IP181/
http://www.ndu.edu/inss/strforum/forum63.html
There are undoubtedly many more, but these are just a few of the links
I have collected. As I said previously I remain to be convinced of the
utility of creating a special category of 'beast' which is a system of
systems. A brief skim through these should give you some idea of what
the term is coming to mean within the defense community in particular.
regards
George
regards
George