[starlogo-users] Activities [on behalf of Irene Lee]
Eric Klopfer
klopfer at MIT.EDU
Fri Nov 16 20:47:10 EST 2007
Here are a few games we use here in Santa Fe New Mexico to illustrate
complex systems in which agents (kids) follow simple instructions and
patterns emerge. One is called "Swords and Shields" another is called
"Walk & Turn". We usually start with "Walk & Turn" because it is
more basic.
In "Walk & Turn" each student plays the role of an agent with a simple
rule. First students are asked to form a large circle then are told
they are to follow a rule: "Turn to face the person on your left
while keeping this heading constant, take three steps." Students try
it once then make predictions about what will happen if they repeat
the rule over and over again. Students conduct the experiment and
test their predictions in real-life. Then we bring students into the
computer lab to look at a computer model of the same activity, Walk &
Turn. Using the computer model as a test-bed, students run
experiments on the model by changing variables such as the number of
agents and the initial configuration (choice of geometric
configurations.) We ask "Does the same pattern emerge regardless of
the settings of the variables?", "Does this model and its outcome
match what we experienced in real-life?". Next, while still in the
computer lab, we introduce the concept of randomly-distributed agents
on the canvas (artificial landscape). We ask what the result of the
simple rule would be if at the start the agents were scattered
randomly rather than arranged in a circle or regular shape. Again, we
make predictions then test those predictions against what the model
produces. Finally, we go back outside and run the experiment in real-
life using the "random" initial configuration. We ask "Does the real-
life outcome match what the model generated? Why or why not?"
(I've attached two versions of the Walk & Turn model, one with scatter
and one without. Thanks to James Taylor of Santa Fe Preparatory
School for developing this activity.)
In "Swords & Shields" participants are asked to secretly select one
person to be their "sword" and a different person to be their
"shield". (participants do not communicate their choice to anyone)
They are told that their objective is to always have their shield
between them and their sword (thus protecting them from the sword.)
When the instructor say "go", use your "Shield" to protect you from
the "Sword." In other words, you must keep the person who is your
"Shield" between you and your "Sword". (***It is important to mark
out an arena with boundaries to play this game within. I'd recommend
moving furniture and desks out of the way.***) Ask for predictions on
what might happen.
Try out the instructions. Discuss what happened and why.
A variation on this game is called "Monkey in the Middle". Again you
secretly select two people but this time your objective is to stay
between the two people you chose. Ask for predictions on what might
happen. Try out the instructions. Discuss what happened and why.
Note that the Walk & Turn models are in StarLogo TNG while the Swords
& Shields model is in StarLogo 2.2.
Enjoy! Let us know how it goes.
--Irene Lee
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Walk&TurnScatter.sltng
Type: application/octet-stream
Size: 13357 bytes
Desc: not available
Url : http://education.mit.edu/pipermail/starlogo-users/attachments/20071116/38818584/attachment-0002.obj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 2. Swords and Shields.slogo
Type: application/octet-stream
Size: 26827 bytes
Desc: not available
Url : http://education.mit.edu/pipermail/starlogo-users/attachments/20071116/38818584/attachment-0001.slogo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Walk&Turn04.sltng
Type: application/octet-stream
Size: 12814 bytes
Desc: not available
Url : http://education.mit.edu/pipermail/starlogo-users/attachments/20071116/38818584/attachment-0003.obj
More information about the starlogo-users
mailing list