Fast demonstrations of complex behaviour

["John Farenden"]

George Richardson wrote:

> ... I cant see what this teaches me about complex systems...

and Niall Palfreyman wrote something like this:

>I cant see how I can disagree with George, even though I felt I should

and Moxnes Erling wrote

>While I felt Georges question was highly appropriate, I see some use for
the "equidistance demonstration" or for any demonstration of complex
behaviour.

I feel a bit out of my depth in this debate, but I used it with two
results....

1. It was a fun ice-breaker to get a group to begin to get to know each
other - I said pick two people you dont know and make yourself
equidistant - at the end of the exercise they had to introduce themselves to
each other. Thats not really complex behaviour is it, but it was fun.

2. After removing one person from the pack and telling everyone to adjust
themselves - which everyone imemdiately thought was no big deal - they ended
up in a state of considerable turbulance for almost as long as it had taken
them to settle down in the first place.

BTW its essential to remove the right person - first time i did it I chose
someone who had had no other people choose him as one of their two
partners - therefore the group remained at equilibrium and the poor chap
developed an inferiority complex !

The lesson for us as a consultancy was that when you are asked by a client
to make what they percieve as a small change to a project, dont immediately
assume that it is a small change and think it through ! If they dont
believe you, try the game on them.

John Farenden
From: "John Farenden" <john.farenden@secta.demon.co.uk>

[Niall Palfreyman]

Moxnes Erling schrieb:

> With many people involved, the task is likely to produce complex
> time patterns of movements. The point can be made that it is impossible to
> predict behaviour without the use of computer simulation.

If this is the extent of the results, then I think I agree with George -
there are simpler ways of illustrating complicated behaviour. The point
of complexity is that complicated behaviour is capable of displaying
certain symmetries ("plexes"), like for example all of my cells go their
own sweet way, but all end up with a cell membrane. The membrane is not
a static symmetry, but a dynamic symmetry, with its contents constantly
changing, but the behavioural symmetry that we call a membrane is robust
across all cells.

So I have a couple of questions for those of you who have run this
demonstration:

1. Do the participants ever come to a standstill?
2. If so, is there any symmetry in their final configuration? Is this
symmetry robust across different runs?
3. If they dont come to a standstill, do they at least exhibit some
symmetry of behaviour, like say milling about in a circle rather than an
amorphous blob? Is _this_ symmetry robust across different runs of the
demonstration?

Id really like to understand this exercise. It seems to me it has great
potential - Id just like to understand what it is.

Niall Palfreyman.
From: Niall Palfreyman <niall.palfreyman@fh-weihenstephan.de>

["Rohan Lulham"]

Good-ay,

Again probably someone out of my depth, but if I may offer one comment
relating to a possible addition to the complex systems exercise.

Would it also be of relevance to ask one or two, or all group members to do
other things such as shake a leg or put their hands in the air, before you
ask one to actually move. This may show that manipulations can be also made
to the system, which from the outside make it look different, but because
the do not relate to the controlled variable do not effect the functioning
of the system.

Similarly, say in a class situation, it may be interesting the week after
doing the exercise, doing it again where a person doesnt move there feet
but leans over to one side. Some people may interpret leaning as a change in
the controlled variable, distance between people, while other may not
creating a more complex response in the system.

These views come from a pretty naïve but developing systems perspective so
hope they are relevant.

Rohan Lulham
PhD Student
Department of Architecture
Uni of Sydney
Australia
From: "Rohan Lulham" <rlul2096@mail.usyd.edu.au>

[Moxnes Erling]

Regarding my reply to Niall Palfreyman about the potential usefulness of the
equidistance demonstration:

My interest in this demonstration was motivated by its simplicity. It only
require participants and some open space. I am not aware of any simpler or
faster demonstrations. The points I made about its usefulness were motivated
by the core of system dynamics and not by the field of highly complex
dynamics. If one wants to carry the demonstration into this field the
following addition could be helpful.

After having run the standard demonstration with the equidistance rule, a
replication to test the forecasting ability, and an equidistance rule with
the addition of distance minimization to see the effect of policy change,
the rule could be changed to: equidistance, distance minimization, and a
dynamic choice of partners such that each person is used by only two others
(that a link is occupied could be indicated by holding hands).

The latter set of rules will produce a pattern and a symmetry that Niall is
searching for. It is important that the outcome is not mentioned by the
organizer. In the debriefing it is pointed out that the local decision rules
lead to the equilibrium pattern, it is not organized by anybody except the
designer of the system (another important point in system dynamics). One
could point to similarities with cell membranes, schooling behaviour etc.
Since the latter variation comes after earlier variations of the decision
rule, the participants will possibly be sufficiently focused on the local
application of the rule that they will not immediately see the solution. It
sounds more trivial and familiar if one just tell the participants to join
hands.

Particularly the third and the fourth variations should be ideal as an ice
breaker in many settings.

Best regards,

Erling Moxnes
From: Moxnes Erling <Erling.Moxnes@snf.no>


Erling Moxnes
SNF, Breiviksveien 40, N-5045 Bergen, Norway
Tel. +47 55959526, Fax. +47 55959439

[Niall Palfreyman]

Bob Eberlein schrieb:

> I found most of the demonstrations discussed to be entertaining and
> attention getting. They all, however, seem to fall short of my
> usability criteria.

I like the usability criterion - Id never really thought about it, but
its certainly a criterion I tend to use in such demos. I guess it
applies to the egg demonstration - I used this to illustrate the simple
fact that a system with twice as many components is not necessarily only
twice as complicated. The demonstration sets up a delayed feedback loop
between the forefingers of the two participants, which then leads to
instability. It would be easy enough to model this situation with two
codependent stocks forefinger_1 and forefinger_2, incorporating a delay
between them, and then to experiment with the delay length to see how
this affects stability. The class could then discuss (and maybe model)
ways in which the delay in the real-world demonstration could be
shortened, and then try out these ways in practice.

Hey, what dyou know? I have a new lesson plan. Thanks for the tip!

Niall Palfreyman.
From: Niall Palfreyman <niall.palfreyman@fh-weihenstephan.de>

[Niall Palfreyman]

Tim Joy suggested that I gather together the various suggestions made
last week of quick "attention grabbers" which demonstrate the need for
taking account of complex behaviour. The following is the list I have
here, and I apologise in advance if I have left any out.

Best wishes,
Niall Palfreyman.

---------------------------------------------------------------------
1. Clyde Jenkins, communicated by Bill Harris:
Take a raw egg thats sitting in a container. Ask one person to pick it
up with their two index (first) fingers and move it from one container
to another.

Then ask a pair of people to do it, with one person contributing the
left hand and the other the right.

Without good feedback between them, theyll have a lot of fun (?!)
trying.

---------------------------------------------------------------------
2. Andy Ford:
I no longer have Andys original text for this, so Ill have to
paraphrase.
He told the story of James Gleick giving a presentation on complexity at
Santa Fe. When he walked onto the stage, he was given a big round of
applause,
for which he expressed his thanks and pleasure, and asked if the
audience
could please repeat the applause, but this time in unison. Without any
further
instructions, the audience proceeded to start clapping and drew
themselves
together into unison. James Gleick then stated that this is a form of
schooling
behaviour, where complexity is handle without the need for an explicit
leader,
and he then used this as the starting point for his lecture.

---------------------------------------------------------------------
3. Ashley Woolmore
Im not sure that this is the sort of thing that you are looking for -
and
may be a bit of a squeeze in the time that you have. I use it for
groups of
about 10-30 depending on the space available.

Essentially it is a demonstration based on CAS and simple rules
generating
complex behaviour. It is also about the consequence of change in one
part
of the system having an impact across the whole.

What I ask people to do is to stand on their feet. They then decide to
fix
on two other people in the group. They keep this decision to
themselves.
The task for the group is to position themselves equidistant from the
two
people that they have picked. They have to move around in order to do
this.
The experience of trying to keep your distance is fun - a bit like
hearding cats!

After a period of time the group will become stationary (at least it has
so
far every time that I have done this demo...). The seemingly random and
difficult to predict moving about of people can be used as an analogy
for a
complex system.

A powerful second step for this demo is then to ask one person only to
change their selected target-people. Everyone else remains with their
old
ones. This person will likely then have to move, which sets a chain
reaction of people moving. The impact of one change has a consequence
for
the whole system.

The whole thing can take just a few minutes. In some situations I have
led
a discussion and de-brief which can take half-an-hour.

---------------------------------------------------------------------
4. Bo Newman:
If you want to extend this experiment (and have the time, the groups
interest/commitment, and have someone with sufficient background
in complexity theory to lead the de-brief ) try the following:

(1) Repeat the process, but this time, without the rest of the group
knowing, instruct one person to change their focus as soon as they
feel they have satisfied the initial requirement. In other words, as
soon
as that person feels they are equal distant from both of their
"targets," they are to drop one of their targets and pick another. And
they are to continue to do so each time they feel they have satisfied
the
equidistance rule.

(2) Repeat the process again, but this time have 2 people changing, then
3
people, etc.

(3) Dont pre-select people to continuously change their targets, but
rather allow everyone to change their target if they feel the target is
acting erratic (moving around too much or with no consistent pattern)
and
trying to achieve their goal (equidistance from two selected targets)
would
be too difficult, or take too long. When the people feel they personally
have achieved the goal, have them raise their hand and keep it up as
long
as they feel they continue to satisfy the goal condition.

---------------------------------------------------------------------
5. Bill Braun:
Some additional variations that might be interesting. A large space is
probably required.

One, have the group divide into four subgroups, each in the corner of a
room. Have one person leave their corner and walk (where and how fast is
up
to them). "Release" another person, then another, etc. until everyone is
"free".

Two, when the group stops have someone at the edge of the group
reposition
her/himself to the opposite side of the group by walking through the
center
of the group. Then by walking around the perimiter of the group.

Three, when the group stops have two people (from anywhere) leave the
group
and move in the direction of their choice. See if the group splits,
remains
cohesive or passes through stages with same or different results.

Compare/contrast the collective behavior of the group based on different
starting conditions and different "course corrections".

From: Niall Palfreyman <niall.palfreyman@fh-weihenstephan.de>

[George Richardson]

Concerning the group milling around exercise in which everyone tries
to stay equidistant from two people that individual privately
selected:

Ive done this several times, and thought about (but not for very
long), and I cant see what this teaches me about complex systems.
Could someone enlighten me? As Jay used to say (maybe still does),
Why would Fortune Magazine write an article about this?

...George
--
-------------------------------------------------------------------
George P. Richardson <gpr@albany.edu>
Chair, Public Administration and Policy Ph: 518-442-5258
Rockefeller College of Public Affairs and Policy Fx: 518-442-5298
University at Albany, Albany, NY 12222 <http://www.albany.edu/~gpr>
-------------------------------------------------------------------

[Niall Palfreyman]

George Richardson schrieb:

> ... I cant see what this teaches me about complex systems...

Well, I must say, Im gobsmacked at the deafening silence in answer to
Georges question. My first reaction was: "What a silly question -
doesnt the man know _anything_ about complexity?" Then I started trying
to think of a good response, and found to my dismay that I couldnt.
After two days of unsuccessfully thinking about it, I realised that I
owe you an apology, George. Ive known of this exercise for a number of
years, and have always sort of assumed that it is a good one without
ever really thinking about it. Just shows you the power of questioning
mental models, doesnt it?

I can only come up with the following answer: I see it as an example of
schooling behaviour, and so of how a simple local rule, when universally
applied, can strongly constrain the complexity of a milling group. I
must confess to never having done the exercise with a group. Does a
particular pattern come out of it, like a circle or something? If so,
then I see it not so much as an exercise in the arising of behavioural
variety, as in the emergence of simplicity in global behaviour.

Im strongly reminded of Ian Stewarts distinction between "simplexity"
and "complicity", and Id love to hear replies to Georges question from
people whove successfully used the exercise.

Niall Palfreyman.
From: Niall Palfreyman <niall.palfreyman@fh-weihenstephan.de>

[George Richardson]

>I wrote:
>
>> ... I cant see what this teaches me about complex systems...
>
>And Niall wrote: ...My first reaction was: "What a silly question -
>doesnt the man know _anything_ about complexity?" Then I started trying
>to think of a good response, and found to my dismay that I couldnt.

The only thing Ive been able to see in the exercise is that a few
simple rules (one, in this case, to stay equidistant from two others)
can generate rather intricate, complex behavior (the milling about).
As I understand it, that idea has been around since at least
Maruyamas "The Second Cybernetics" article (1963) and Conways game
of "Life" and now finds expression in the few rules found sufficient
to produce rudimentary schooling behavior in fish and flocking
behavior in birds.

Is that the "IT" people are supposed to get from this exercise? And
if thats it, then what do we do with it?

...George
--
-------------------------------------------------------------------
George P. Richardson <gpr@albany.edu>
Chair, Public Administration and Policy Ph: 518-442-5258
Rockefeller College of Public Affairs and Policy Fx: 518-442-5298
University at Albany, Albany, NY 12222 <http://www.albany.edu/~gpr>
-------------------------------------------------------------------

[Moxnes Erling]

George Richardson wrote:

> ... I cant see what this teaches me about complex systems...

and Niall Palfreyman wrote something like this:

I cant see how I can disagree with George, even though I felt I should

While I felt Georges question was highly appropriate, I see some use for
the "equidistance demonstration" or for any demonstration of complex
behaviour. The beauty of the "equidistance demonstration" is its simplicity.

1. Each actor can be modelled as operating with a simple feedback loop where
equidistance is the goal and where walking in an appropriate direction is
the policy. With many people involved, the task is likely to produce complex
time patterns of movements. The point can be made that it is impossible to
predict behaviour without the use of computer simulation. And, that the
system "solves" this set of coupled differential equations by itself,
without any coordination. The fact that all actors face identical tasks is
not important from my point of view, in the debriefing one could simply
comment that in other systems task descriptions differ (e.g. buyers and
sellers as in typical SD models). It seems obvious that complexity is
maintained when rules are allowed to differ.

2. One could ask people to do the demonstration twice, each time starting
from the same initial positions, using the same decision rules, and each
time noting the final positions. The first demonstration could be seen as a
forecast of the second. Probably small inaccuracies in peoples actions
would lead to surprising (?) deviations from the predicted outcome (the
demonstration does not have a unique equilibrium). If so, one could make the
point that it is difficult to make accurate forecasts.

3. By changing the rules to equidistant and mimimum distance, I guess the
demonstration would tend towards one unique and very intimate equilibrium.
(Is there some general tendency here that mini- or maximizing behaviour
among the participants tend to weed out equilibria??). The intimate
equilibrium could be predicted by simulation (although it is cumbersome to
explicitly model this system). Points 2 and 3 can be problematic if they
induce the participants to think that equilibrium and prediction are central
goals of simulation. Thus in debriefings it is important to stress that the
purpose of simulation is to find appropriate policy rules. If the intimate
equilibrium is perceived as the problem to be avoided or as the desired
situation to be in, policy rules should be chosen accordingly.


Best regards

Erling Moxnes


Erling Moxnes
From: Moxnes Erling <Erling.Moxnes@snf.no>
SNF, Breiviksveien 40, N-5045 Bergen, Norway
Tel. +47 55959526, Fax. +47 55959439

[Bill Braun]

At 10:33 AM 4/30/01 -0400, George Richardson wrote:
>Is that the "IT" people are supposed to get from this exercise? And
>if thats it, then what do we do with it?

Perhaps the value varies with the experience and knowledge of the people
engaged in the exercise. I have not tried this exercise yet but I plan to.
For a group of students who have never asked, "So, what is a system?", I
anticipate it will be an effective catalyst to begin pondering that
question. It is an effective introduction to how rules/structure (of the
structure > patterns > events pyramid) affects system behavior.

It is a good demonstration that leads to some learning insights, such as,
"You cant do just one thing" and "The behavior of any part of a system is
interdependent with the behavior of at least one other part of the system"
and "No part has an independent affect on a system."

I know for the students with whom I work this should be a fun and effective
means of revealing some mental models and suggesting that they set them
aside as a prelude to learning.

Bill Braun
From: Bill Braun <medprac@hlthsys.com>

["Jack Ring"]

One thing a practitioner might get from "IT" is the great importance of
Goal setting for humans involved in networks (and arent we all). I
think this applies directly to the Tragedy of the Commons which is best
resolved by "installing" a cascade from (common) Purpose to (applicable)
Principles to (individual) goals in all participants. Such systemic
control is likely the only cost-effective form that works.

The Northwest Atlantic Marine Alliance, www.namanet.org, is an example.

One mathematical foundation for this can be found in
http://www.sie.arizona.edu/sysengr/wymore/optimal.html in which A. Wayne
Wymore, Prof. Emeritus, Systems Engineering, shows that, contrary to
folklore, an optimal system can be achieved by combining optimal
subsystems if there are certain restrictions on the tradeoff function.

In the Fast demonstration every participant has the same goal.

To further exemplify CAS, other demonstrations could include
individualized goals (as in mass customization) to show that some goal
sets lead to stability while others do not.

Jack Ring
From: "Jack Ring" <jring@amug.org>

[Elise Axelrad Weaver]

>3. If they dont come to a standstill, do they at least exhibit some
>symmetry of behaviour, like say milling about in a circle rather than an
>amorphous blob? Is _this_ symmetry robust across different runs of the
>demonstration?

I was in a class where we tried the exercise.

We did come to a standstill with one symmetry: A number of us chose the
same 2 people as our foci, because the two students often sit together.
Nobody stood directly between them, but two clusters of about 4 people
each emerged, one to either side of the line which joined these two students.

cluster1

focal student 1 focal student 2

cluster2





Elise Axelrad Weaver, Ph.D.
Research Assistant Professor
Center for Policy Research
Milne 300E, University at Albany
135 Western Ave.
Albany, NY 12222
(518) 442-3864; eaweaver@csc.albany.edu

[Bob Eberlein]

Hi Everyone,

I want to make a comment on this discussion and it goes back to
something stated quite innocently by someone some time back that he or
she had tried to model a particular exercise but had not been able to
make any headway. I do find it troubling that we are finding quick
presentations that are not really that easy to model in a nice compact
system dynamics model. As a student I think I would be pretty uneasy
taking a course teaching skills that were of no use in dealing with the
motivating problem introduced during the first lecture.

I actually have this problem somewhat even with the beer game, which is
pretty easily modeled in a standard system dynamics format. I have
never seen anyone, at the end of the course, go back and say lets build
and analyze the model for this and then go back and restructure the
business rules then play the game again and compare notes. Or, even
more of a challenge, just play again with the same rules and try to do
better. If the lesson is meant to be that the structure determines the
behavior and there is no room for improvement that is a dismal lesson
indeed.

I found most of the demonstrations discussed to be entertaining and
attention getting. They all, however, seem to fall short of my
usability criteria.

Bob Eberlein
bob@vensim.com

[Bill Harris]

Bob Eberlein wrote:

> better. If the lesson is meant to be that the structure determines the
> behavior and there is no room for improvement that is a dismal lesson
> indeed.

You should have seen the version Ive taught.

Seriously, I was asked to teach it to a group about 5 years ago, and I
did it pretty much by the book: a bit of talking (lecture), a dash of
the PBS video, and one run through the game (interesting, because I was
"teaching" a bunch of manufacturing managers; they could see what was
coming--mountains of beer--but they seemed powerless to stop it). When
we finished, my co-facilitator and I became convinced we had taught one
thing extremely well: personal incompetence.

That not quite being our goal, I tried to restructure the class.

I took as a goal helping them understand that some simple looking
problems were indeed complex and that these problems could be best
solved with the aid of testing (simulation). When we taught it the next
time, we compressed the standard part of the course into the morning.
By lunch, when people were laughing at the stacks of pennies (beer) in
front of them or bemoaning their lack, they were also saying, "Well,
thats only because you gave us these funny rules we had to use." We
said, "Okay. Heres the second part. We want you to make up your own
rules. You have to use the boards we gave you, but everything else is
up for grabs."

They took about an hour, I think, to redefine the rules (policies), and
then we started a new simulation.

Howd it go? Worse than the first! Even though these were smart
people, experienced in manufacturing, they ended up with more overstocks
and outages than they did with our rules.

We closed with a discussion about the value of testing such "process
improvements" before subjecting the organization to them, and we
referenced two options: manual game playing, as we had done in the Beer
Game, and automated simulation, as we projected from a laptop running an
SD model of the Beer Game and ran some experiments to show the effects
of different strategies.

While it wasnt perfect, I think this approach came closer to meeting
our goal of showing them what they could do, rather than only impressing
them with what they couldnt do.

If I get time, I may write this up and post it on my Web site.

Bill
From: Bill Harris <bill_harris@facilitatedsystems.com>
--
Bill Harris 3217 102nd Place SE
Facilitated Systems Everett, WA 98208 USA
http://facilitatedsystems.com/ phone: +1 425 337-5541