Subject:
|
Re: multi-RCX dreaming
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 17 Jun 1999 20:36:44 GMT
|
Original-From:
|
Luis Villa <LIV@DUKE.EDUihatespam>
|
Viewed:
|
943 times
|
| |
|
|
On Thu, 17 Jun 1999, Mark Crosbie wrote:
>
> In message <9C998CDFE027D211B61300A0C9CF9AB46F22C1@SNYC11309>, "Stalker-Wilde,
> Graham" writes:
> > get them to play iterative prisoners' dilemma.
> > How is up to you. it could be a pure programming task - with IR the only
> > communication, or it could be more physical...
>
> My thoughts exactly. Though I was leaning more towards experimenting with
> different Evolutionary Stable Strategies (ESS). See if an RCX using
> tit-for-tat ESS can invade a population of RCXs using a different, more
> greedy, ESS. See Robert Axelrod's book for more details: "The Evolution of
> Cooperation".
I'm a poli-sci major (in my spare time :) and own a copy of the book as a
result of that. I can't believe I didn't think of that myself.
> My interests tend to lean towards emergent cooperation: can 6 RCXs do
> something together as a team that one RCX cannot do?
Those are the kinds of things I was thinking about. A simple demo that I
have heard suggested elsewhere is to scatter a bunch of balls on a
surface. Each robot has basically five instructions:
1) find a ball.
2) pick it up.
3) find another ball.
4) put down the original ball next to the new found ball.
5) repeat (starting in such a way that the ball found in 1' is not the ball
found in 3.)
Eventually, this will collect all the balls in one location.
> Another interesting task: program "flocking" behaviour. The rules that govern
> schools of fish or flocks of birds are deceptively simple. Sort of "Stay a
> constant distance from your nearest neighbour" and "Always head towards the
> perceived point of greatest density of neighbours". These two rules allow a
> flock/school to stay together, avoid obstacles and still regroup into a flock
> as they fly/swim around things.
Good idea! I remember reading some articles about this some time ago.
I'll have to look it up and see if I can find some...
> So what is the collective noun for a group of RCXs? A "brick" of RCXs :-)
Hehe... what is the collective noun for a group of politicians? A "lack"
of principles...
> The fun part is that you can tie this in with a lecture on
> Evolutionary Computing and Emergent Cooperation.
The thought had crossed my mind :) I'm only vaguely responsible for
lectures and topics (my main job is "make the lego do something with our
techniques") but nothing is set in stone.
-Luis
#######################################################################
Profanity is the one language that all programmers understand.
-Anonymous
#######################################################################
--
Did you check the web site first?: http://www.crynwr.com/lego-robotics
|
|
Message has 1 Reply:
 | | Re: multi-RCX dreaming
|
| ... (...) this i disagree with, unless the balls can be marked somehow, another bot has the possibility to pick up the one that was placed in the proper place, now, if you were to mark a corral and have the balls placed in that it would stop that (...) (25 years ago, 8-Jul-99, to lugnet.robotics)
|
Message is in Reply To:
| | Re: multi-RCX dreaming
|
| (...) My thoughts exactly. Though I was leaning more towards experimenting with different Evolutionary Stable Strategies (ESS). See if an RCX using tit-for-tat ESS can invade a population of RCXs using a different, more greedy, ESS. See Robert (...) (25 years ago, 17-Jun-99, to lugnet.robotics)
|
4 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
