Slashdot Mirror
Simulating Societies
blamanj writes "Most of us were exposed fairly early to Conway's game of Life.
A few simple rules produce a fascinating variety of behavior. Now, it
appears that similar simulations can predict the behavior of populations and human societies."
Psychohistory!
"Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
A couple more samples...
Music created using the game of life
Geez, if this mentioned globalism and post-911 trauma, I might start thinking it's a Katz article from the way it reads.
Dude, get a life.
I had to write a "Life" program for the Pr1me as part of a college project years ago. It was ok when run on a VDU, but some fool ran it on a teletype... one box of paper later.. it was turned off.
It might be profitable for certain companies to monitor new Slashdot stories as they relate to human behavior. The rate of influx for new stories is bound to be inversely related to the workers' productivity :)
If you're into this stuff, this link is cool.
Wow -- this guy really has some issues; perhaps he doesn't understand the concept of "if you don't like it, opt out.."?? Better yet, stop your bitching and start your own site, mod your own posts, and then complain about its inherent flaws. Some people may have way too much time and education -- this is an example....
...we are from the government - we are here to help...
I don't think simulations are ever going to get it right, because of so many possibilities that each of us encounter. People are too wishy washy, same events effect people differently, etc etc.
A slip of the foot you may soon recover, but a slip of the tongue you may never get over. -Benjamin Franklin
I predict that it your screensaver is Life, you'll get no work done.
Preachers (albeit self-inflated ones), Theologians, Prophets and madmen have been doing that for years, albeit with little success.
The primary problem is that the raw data cannot predict the movement of society, so therefor conjecture must be used. The conjecture is based on a hypothesis which is based on one of the obove basic viewpoints: religion vs. lack-thereof, pessimissm vs. optimism and basic intelligence of the average human vs. lack-thereof.
Unless the person who writes the simulation is a prophet or exceptionally gifted, the software will be as flawed as any other model.
My $0.02 will always be worth more than your â0.02, so
This article is really describing modeling using multiagent systems. Though very simple multiagent systems may resemble cellular automata (such as Conway's Life), they are not the same thing. Though they have been described in very convenient graphical representations using grids in the article, agents can model more complex behavior and need not be determinisitic (i.e. they may have a random element).
Another way to look at it is that cellular automata like Life use a single deterministic rule to govern the whole system. Agent-based systems, on the other hand, model goal-oriented behavior of the individual objects.
Again, Conway's game can be viewed as a very special case of an simple agent system, but the spirit of what is being done with agent systems is typically more involved. Comparing these systems to Conway's game of Life may create an incorrect impression for those not familiar with agent programming.
If you found this article interesting, their book is a great exposition of their early work with emergent behaviors. You can find it at Amazon here:
Growing Artificial Societies
There is a similar article on complexity and emergent behavior in the latest Harvard Business Review.
-XDG
Hmmm... So the simulation is accurate, but I would hypothesize that it does not show that a free society will trend towards "honesty."
Nonperiodic Central Trajectory
Remember Asimov's Foundation series where Dr Seldon used mathematics and psychologists to predict and model the behavior of populations and human societies.
Asimov has a habit of predicting scientic advances such as robotics(Everyone know Asimovs laws of robotics ?)
Ok he was basing it on the presumption that you could predict the behviour of very large population (ie whole planets),but the concept was the same
Better watch out for the Mule...
does any1 think we'll stop pretending that we're NOT on the brink of coolapps, & that fud is not dead?
nt
Someone should run an experiment and force traffic to be out of my way on my way to and from work. That would make me honestly less pissed when I get there. I would then be less corrupt and more apt to leave the office supplies at the office.
one two three four five ?!! That's the combination on my luggage!
The "Foundation" series by Issac Azimov never really seemed too far fetched to me. The ability of dedicated mathematicians to predict the course of large enough groups of human beings seemed to me to be perfectly reasonable, given enough variables and a population size that minimizes the chance for really unique/aberant behaviours. Now we have the computing power to back it all up.
For those of you who will counter that I'm neglecting the point of the Second Foundation manipulating things... don't spoil it for me. Seldon still had to get at least the first several decades right you know.
I miss Asimov. I read almost all his sci-fi stories (except of the short stories).
There's one book of him I haven't read yet because I couldn't find it: A Pebble In The Sky.. I wonder if I can find it anywhere.
- Elkobim
I want tender love now!
Elkobim
One thing that I've found interesting is how closely *any* group can be predicted -- this from the three or four required sociology courses in college. Many of use here pride ourselves on having different values than the mainstream population. However, the behaviour of the niche groups can be eerily predicted by statistical models to the point that it's now a business tool and not just cool science. So we may not be able to predict that an individual is a devoted Bob Dylan fan, but they can probably see upswings in folk music and tie dyes whenever a war is brewing in the (insert region here).
--
Everybody must get stoned.
Again, Science fiction predicts fact!
Isaac Asimov, in his wildly popular "Foundation" series (read it if you haven't), predicted that eventually human actions as a group (not individuals) could be calculated through computer simulations. He called it "Psychohistory" (as previously noted). In the novels it was a matter of probability, almost certain probability in most cases, butunlikely events still were able to mess it up. (For example, a telepathic genetic abnormailty, or "Mule," was born, and using his powers basically conquered the galaxy, thereby screwing up predictions)
Anyways, its really nifty how many cool technologies have been predicted by science fiction authors.
Other large scale societal modeling took place with The Club of Rome's Limits to Growth -- It used the SIMULA simulation language to investigate such questions as population growth, resource usage, environmental degradation and capital investment as co-related variables. They came to some very interesting (and even disturbing) conclusions.
Sometimes boldness is in fashion. Sometimes only the brave will be bold.
Unfortunately for people who maintain that man is ineffable and that God is unknowable, the facts are that man is statistically predictable, easily manipulatable and, while he is imbued with a lab animal's right to do whatever he damn well chooses in a carefully controlled experiment, he rarely does so he is reducible to a mathematical theorem.
As for God, when he calls you on the phone, tells you where Bin Laden's hiding and what the results of tomorrow's lotto pick, then you can publish a paper on his existence. Until then, less God and more functioning brain cells, please.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
For those who don't get subtil things (or just happened to miss this one)
When you perdict something people tend to act on the perdiction. Thus God sent a profit to warn Niniva of coming doom, but the people repented and so God no longer needed to send that doom. So does the fact that the people lived (for 100 years before some other country invaded) mean that God doesn't exist, or that repenting will save your life?
If everyone knew the terrorist were going to fly a plane into the world trade center in september nobody would have been there. (other than press, and some engineers to study the situation). If the terrorist knew they were discovered like that odds are they would call the whole thing off, and everyone would then laugh at those who gave a warning about something that never happened.
2000 is a perfect example. There were big comptuer problems related to the roll over from 1999 to 2000, but because there was warning the problems were fixed, so there were no problems, so the warnings must have been uneeded right?
There are many more examples that can be thought of. The point is clear though: warnings are a double edged sword.
However I'm willing to perdict the next terrorist bombing will be in Iseral/Palistine. You are now warned. (too bad I can't be more specific, this will do you little good if you live in that area)
Damn, when I read the header, I really thought we were all talking about The Game of Life
Was I the only one who thought that?
Blah Blah Blah.
What about Black and White. Wouldn't that be a society?
"Power corrupts. PowerPoint corrupts absolutely."
It's not just that the simulations use
simple rules, but we humans use simple rules
too because we are simple minded and are usually
driven by simple heuristics. It's not suprising
that the simulated behaviour closely matches
real behaviour. Fot it to be otherwise would
take a level of intelligence we don't seem
to have.
For those who don't get subtil things (or just happened to miss this one)
Subtle.
When you perdict something people tend to act on the perdiction. Thus God sent a profit to warn Niniva of coming doom,
Predict. Prediction. Prophet. Nineveh.
2000 is a perfect example. There were big comptuer problems related to the roll over from 1999 to 2000, but because there
Computer. Rollover.
was warning the problems were fixed, so there were no problems, so the warnings must have been uneeded right?
Unneeded.
However I'm willing to perdict the next terrorist bombing will be in Iseral/Palistine. You are now warned.
Predict. Israel. Palestine.
Your score is 11. Your rating is JeffK. Thank you for playing!
All employees must wash hands before seeking equitable relief.
From the article:
As in real life, a few A-firms live and thrive for generations, but most are evanescent, and now and then a really big one collapses despite having been stable for years. Sometimes the addition of one slacker too many can push a seemingly solid firm into instability and fission; but you can't be sure in advance which firm will crumble, or when.
Sound familiar to anyone?
The problem with trying to predict sociological events beyond simple "self-segregation" and the like, is there are just too many factors involved.
Imagine trying to predict an event when there are _milliions_ of factors, and every single one has an influence, with each having largely unpredictabily varying degrees of possibility of influencing the overall picture. That is what it is like trying to get a computer to predict sociology.
I could write heaps about this, but basically, I'll say that I don't believe using computers for this sort of thing is useful for anything beyond the blatantly obvious.
This was really cool reading!
I must say I enjoyed this story very much.
Life is too short, die now!
can be found on the other site:
5 53
http://www.kuro5hin.org/story/2002/4/1/55731/68
where this link/story was posted 10 days ago.
Here is a paper (by the same author) on the simulation of the evolution of communication based on kin structure under the Prisoner's Dilemma.
Seastead this.
It is no coincidence that some of the best Cinematography of the early 20th Century came out of the National Socialist propaganda machine.
A couple of years ago, A&E did a glowing, kiss ass, Biography on the woman who made "Triumph of the Will," probrably the most notorious of the Nazi propaganda films.
It should come to no surprise that A&E is owned by Disney.
The article was therefore incorrectly moderated.
I'm pretty sure I had a version of this simulator on my Commodote 64 back in the mid 1980s. Of course, with a 320x200 grid and a 1MHz processor, it took many hours for the segregation to be complete. I remember being fascinated by it.
Could he send one to me please, I'm a little hard up at the moment and I can't predict any coming this bway soon given the current IT situation!!!
Sim City anyone?
I agree that these guys would sound like neo-facist pinheads if they simply said "all we have to do is get rid of all the bad guys and society will become perfect." But they wrote a computer program to make that statement for them, with colored lights and everything. Computers take the rules they are given (like the fact that only corrupt people get arrested, and arrested people are always rehabilitated) and follow these rules to their logical conclusion. How can you possibly argue with that?
The major problems with the models is that they are not very good at handling technological change that in turn makes fundamental changes in the values the models use to make its predictions.
For example, let's say your population growth model includes a value for "food value produced per acre of land". If something comes along that allows more food to be produced per acre, then that'll skew the models to hell.
This actually happened. A new strain of wheat (?) was produced a few years ago that was able to survive in much tougher conditions, and that single-handedly staved off starvation in India.
The same with waste levels. recycling has become much more prevelent, and modern cars are so much better that they're actually starting to _clean_ the air that passes through them.
The models were accurate the day they were published, but the run conditions have changed since.
DG
Want to learn about race cars? Read my Book
From what the article describes, the people doing these experiments have got their research backwatds. Specficially finding that a particular set of assumptions to a simulation generates a result 'like' human society is meaningless unless you also show that the assumptions are legitimate. The racism example was particularly egregious; nowhere is it explained why ignoring the effect of income distribution, access to jobs, the actions of the government, etc on where people lived was valid. It gives the strong impression that showing that racial division arises from inscrutable preferences is attractive for political reasons more than anything.
In the abstract he states:
"In the Demographic Prisoner's Dilemma, neither assumption is made: agents with finite vision move to random sites on a lattice and play a fixed culturally-inherited zero-memory strategy of cooperate (C) or defect (D) against neighbors."
After his citation of Michael Oliphant's paper (1994) Evolving Cooperation in the Non-Iterated Prisoner's Dilemma: The Importance of Spatial Organization published in Brooks, R. and Maes, P. (eds.) Proceedings of the 4th Artificial Life Workshop, pp 349-352, The MIT Press. Epstein proceeds to attempt to justify his paper in comparison to Oliphant's genetic-algorithm paper by emphasizing his definition "culturally-inherited" as follows:
"Perhaps it is worth emphasizing that, in adopting this assumption of a fixed agent strategy, we are not claiming that human strategies are literally hard-wired genetically. Rather, for modelling purposes, we are assuming that they are culturally transmitted from parents to children--vertically transmitted--with high fidelity, like certain religious or ethnic affiliations, tastes, and native tongues. 19 Below we consider the effect of degradation (mutation) in this vertical transmission fidelity."
This definition, as well as from his other descriptions of his algorithms differ in no way from Oliphant's 'genetic' tendencies to defect or cooperate, except to make the environment 2 dimensional instead of one dimensional and to make spatial structure evolve out of variation in "sight" rather than a simple gaussian distribution of mating -- neither of which can be used to distinguish "culturally-inherited" from "genetically-inherited" traits.
While it is interesting to extend Oliphant's work on genetic algorithms to 2 dimensions, it sheds little new light on the subject.
What would have been far more interesting, especially from the Brookings Institute's charter, and from Epstein's position of responsibility for defense policy analysis there, would have been to do a genuine investigation of cultural transmission in the presence of genetic selection as well as cultural selection:
Then study under what conditions genotypes arise that tend to transmit 'cooperator culture' while they, themselves, transmit 'defector genes'.
The above extensions to Oliphant's one dimensional gaussian model should be sufficient to illuminate the nature of such 'meta-defection', although I'm sure variations and elaborations on his minimalist environmental model would become obviously interesting in short order.
Seastead this.
So you develop a model of human populations that works moderatly well. It won't give you an exact image of what is going next but still follows the same trends. You modify it slightly in some real world way. Like as in this example allowing the peace keepers more mobility in arresting people that commit violence. It is doubtfull that it will work perfectly. But it might reduce the chance that an act of genocide occurs. So in essance a probabilistic model of the world to which modification has a probabilistic effect on the real world. That is very usefull. Sure it won't work in all cases but it's much better than nothing. It reduces the risk of a viable solution being canned because it just happend to run in to a set of random events that where detrimental. Where as the normal case for this solution is quite effective. Besides your arguments are assuming absolutes. Process A fails once so it is as bad as process B which fails once. However when you look at the two one model gets things right 99% of the time and the other gets it right 99.999 percent of the time. Which are of course not equally flawed. To illustrait with a real world example think of weather predictions. Over the last century we have improved the length of time that our predictions are adequatly acurate from a few hours to 5 days. Both of which are based on raw data. Of course after a period of time your guess is as good as environment canada. Except that as time goes on we refine our weather model and update the data to a current state. Allowing probabalistic accuracy for another period of time. So if we can do weather why can't we do society? What says we can't predict free will? After all we do it all the time. Think about the predictions we use about other drivers when we are driving. How about how people react in a social setting. There are limits on humans normal reactions so why can't we try to model it with a reasionably usefull accuracy.
Nobody suspects the butterfly!
Quote from the article:
``Neither she nor anyone else, Epstein included, believes that an array of little dots explains the Rwandan cataclysm or any other real-world event; the very notion is silly. What the simulation did suggest to Des Forges is that disparate social breakdowns, in widely separated parts of the world, may have common dynamics--linking Rwanda, for instance, to other horrors far away.''
Well thats not what the model is at all. First the people do not get rehabilitated. They are just afraid of getting caught. When they are past a certain level of fear they start turning other people in. Thats not a police state thats a snitch state. However there are alot of flaws when compared to real world situations. No "this worked before" events. No loyalty. No exclusive dealings with friends. No lying to the authorities about a deal. No believing one person over the other. Lots of problems which translate into real societies shifting more rapidly and more towards either direction than the model.
Nobody suspects the butterfly!
Does anybody know the iterative rules for the repeated prisoner's dillema simulation? (The one with George Washingtons).
I've toyed around but been unable to get them.
Thanks in advance
-Knots
Anarchy$ dd if=/dev/random of=~/.signature bs=120 count=1
The article reminded me of the old story of the experimentall physicist who runs excitedly up to his theorist colleague, exclaiming "Look! I can show that A > B!" The theorist says, "That's easy to explain. [Explanation deleted...]" The experimentalist says, "Did I say A > B? I meant B > A.", to which the theorist replies, "Oh, that's even easier to explain."
The models described seem far too simple to describe something as complicated as society. As a physicist who has dabbled in biology, I know the perils of applying simple models to biological systems. How sensitive are these models to the addition of another type of interaction between people, or another outside influence? For every simple model that shows A>B, I can come up with one that shows B>A, unless the simple model is very well rooted in fhe fundamental physics (or sociology) of the problem. I don't believe that the fundamentals of sociology are well enough established to make these models believable.
For example, consider the Schelling model of segregation discussed in the article. From a physicist's point of view, this is a statisictal simulation of a system of two types of particles on a lattice, with an attractive interaction between particles of the same type. There's no temperature, so the system will phase separate, since that's the lowest energy state. No surprise there. A five minute chat with a physicist could have saved Schelling a lot of computer time. The more interesting question is what happens when you add some randomness in the form of temperature. Then the system will phase separate below a certain temperature, and form a single mixed phase above that temperature. What is the sociological analog of temperature? (Ok, I know that one... If a particle of one type is hot for a particle of another type, then you get particles of mixed type....)
The simulations are cute and I'm sure they're fun to play with, but I wouldn't put much stock in them.
-- Steve
Lost: one sig, witty, 120 chars, sentimental value. Reward offered.
A Multipurpose simulator with programmable rules called Starlogo is available at Starlogo.org
Also read "Emergence" - that book must have been reviewed on Slashdot long ago - same concepts.
-.Shaun.-
Take a look at this cool online journal
The Journal of Artificial Societies and Social Simulations
Which is an interesting result. It suggests that this is the key thing for society to concentrate on in order to prevent disaster.
Of course you don't want to embark on such a course based purely on Limits to Growth, but the value of such simulations is that they tell you where the hidden levers are, even if they can't give precise predictions about what happens when you pull them.
Paul.
You are lost in a twisty maze of little standards, all different.
The author of the article goes out of his way to note that agent-based modelling does NOT predict human society. At the moment, it's used to create analogous models of certain specific traits. The difference (that is not modelled) between an agent in these simulations and an individual in society is that the individual is (allegedly) rational: s/he has the ability to reason about the surrounding and make decisions appropriately, while agents follow extremely simple rulesets.
That said, it's still a pretty cool field. Check out http://www.swarm.org for a GPLed agent-based modelling package.
The program that ended up as the most successful was also the simplest. University of Toronto Game Theorist Anatol Rappaport had submitted a program he called tit for tat. Tit for tat initially cooperated with all the other players. In subsequent turns if the other player it was interacting with had defected last turn, it defected this turn. If the other player had cooperated last turn it cooperated this turn.
Yes, the interactions between people are very complicated, and this game is very simple. Still food for thought though.
For those interested in the subject of simulating artifical societies in silico i strongly recommend:
(Sorry, i'm against linking to online book stores)
Growing Artificial Societies - Social Science from the Bottom Up
Joshua M. Epstein & Robert Axtell
ISBN 0-262-55025-3
So are there twice as many trolls as offtopics, 10 times as many trolls as insightful posts, and so on?
-- Two men say they're Jesus. One of them must be wrong. - Dire Straits
First, the book is full of examples, but nowhere to Epstein and Axtell give you enough information to actually reproduce their results (a classic mark of shady science).
Second, there are parts of the book where they draw conclusions from things that are obviously simulation artifacts (ie. if you change the grid size, these effects disappear or are mitigated severely).
Did I mention their lack of understanding of basic computer science issues? (Their formal training is in the social sciences).
For a pair of scholars at the esteemed Brookings Institute, you would would expect more. Unfortunately, you wouldn't get it.
Don't buy their book.
But that then changes the whole psychohistory bent and Foundation away from predicting the future, and instead takes it into controlling the future. It kinda makes one wonder how much 'nudging' the second Foundation had been doing up to the point where the Mule showed up.
Hmm... and I guess that changes Harry Seldon from a visionary into a Despot. Well, at least he was a benevolent one.
The racism example was particularly egregious; nowhere is it explained why ignoring the effect of income distribution, access to jobs, the actions of the government, etc on where people lived was valid.
No explanation is necessary: if the model accurately predicts how real-world societies work, without using the factors you seem to need to be taken into account, ipso facto those factors are irrelevant. You seem to think the function of research is to support pre-decided goals; thus you have things backwards. In any event, the interesting thing illustrated was that non-racist behaviour by the atoms (even an atom simply wanting to be next to one other of its racial group) still gave rise to racist-seeming patterns. Not "inscrutable preferences" at all.
I'm suprised that no one has submitted the obligatory comment that Bill Gates is the Mule. He sure looks like a mutant. Particularly since Slashdot gave him that nifty Borg-esque makeover.
We need to set up a website, www.WhoIsTheMule.com, and take polls as to who we think is most likely to take over the world.
I'm submitting my vote for Jon Katz.
Used on amazon from $2.00..
3 5 [amazon.com]
http://www.amazon.com/exec/obidos/ASIN/03453356
It is quite a good story, actually.
Right now the Writers of America are boycotting Amazon. Every time you buy a used book from them the author gets nothing, nada, not a cent.
They are the pirates of our generation, the RIAA of the MP3 world.
As with music, where you should buy the CD from the musicians instead of thru RIAA (hint - they make $5 for a $6 CD they sell in person, and $0.02 for a $15 CD you buy thru RIAA) - for books you should buy from the author (e.g. printed book). they get no money for their work when you buy it used.
Note that libraries do kick back to authors - and in Canada and the EU they kick back a big chunk of change. So please check it out at the library before you buy it used from Amazon.
[note - I'm biased, I've sold stories myself]
-
--- Will in Seattle - What are you doing to fight the War?
The Mayans of Central America also disappeared from the area, leaving their cities abandoned.
Minix en español! http://www.es-minix.org
A model pictures reality by simplifying it. This research is emulation.... The problem is that everyone KNOWS the outcomes. The IMPORTANT question is whether the underlying assumptions catch the true causal forces.
And there are real problems with this school of thought, not the least of which is its claim that getting complex interactions out of simple assumptions is any harder than getting complex interactions out of a great deal of assumptions. It should be self-evident that complexity in this type of research stems largely from the number of actors, not the determinants of their behavior.
Deeper problems include assumptions of rationality and intentionality on the part of actors. There is also a tendency towards selection bias and selectivity THAT IS NEVER ADDRESSED. IE, this author may think he explains ethnic genocide in Rwanda, but never points out that his logic fails miserable in places like Switzerland, Brazil, Mexico, Russia and much of the Middle-East, where his model would predict much MORE conflict than we see.
There's a book called In the Country of the Blind by Michael Flynn which is driven by a related premise. Essentially, a group in the early nineteenth century develops a "science of history" that allows them to predict the future (initially using Babbage's difference engine) and, of course, a secret society develops that begins engineering the future. Unfortunately, for me, the book falls into the category of Interesting Premise/Uninteresting Writing.
If you use the data say 1970 - 1985 for building your model, you can use 1986 - 2002 as the 'future' you are trying to predict. Commonly your data is partitioned into a training set and a validation set, usually on an arbitrary basis rather than on the later/earlier basis suggested above.
Be Free: Free Software Tuition
The Journal of Artificial Societies and Social Simulation (JASSS). It is on-line. It is free. It is great.
Belief is the currency of delusion.
People trading stocks for profit is a subset of society. No one has been able to predict market futures accurately. Ask Long Term Capital, the hedge firm full of Nobel economists, that almost took out the world economy four years ago.
The problem is that once someone figures out some new profitable information about the market, it works for a while until enough people figure out the same method. Then it becomes useless.
I expect prediction of society as a whole to fail for the same reason. When people learn what is being predicted, they'll do sometime new and unpredictable.
The whole agent-based modeling paradigm has had a profound influence on social science research in recent years.
.)
Essentially, advances in computing power (grid computing, graphics, CPUs, etc.), theory (complexity studies, systems science, geographic information science) and object-oriented programming paradigms now allow us to simulate cities with incredible levels of detail, often at the scale of individual people, homes, cars, etc. and in time-scales sometimes approximating "real" time. And there is a lot of data out there at this scale to "feed" these models. We're doing SimCity for real! (Well, kind of anyway...
In the field of geography and urban planning, people such as myself are working on these style simulations to help us develop virtual laboratories that serve as artificial cities to test all sorts of hypotheses and ideas about various urban systems: suburban sprawl in my case (http://www.casa.ucl.ac.uk/sprawl), but also traffic congestion (http://transims.tsasa.lanl.gov/), pedestrian shopping in downtown areas (http://www.casa.ucl.ac.uk/streets.pdf), residential location decisions (http://www.casa.ucl.ac.uk/paper32.pdf), etc.
I have some more info. and links at http://www.geosimulation.com if anybody is interested.
MEEP! MEEP!
...who didn't see that coming. :)
Technoli
We could play "The World" in real-time on a huge, distributed network of some kind, something like a mix of E-Bay, Everquest and IRC only much, much greater. Add some CNN Online for thrills and feed /. streams at random. Something like that. Make it browserbased.
We could "simulate" all sorts of events, you know, terrorist attacks, meteor impacts or natural disasters. Anything. The winners would sweep the stakes according to some sort of victory resolution scheme. Maybe THAT could be coded in Perl.
All players could "initiate" actions at any time that would, eventually, over many turns, determine the final outcome. Players could interact with one another according to some proximity scheme. Players could coorperate toward common goals.
At intervals we could make tournaments, where the winners of the local series would compete in the World Series. The World Champion would collect a huge prize and maybe move into The White House.
Hmmm. I think I'll go to the pub...
For the sufficiently clueless, even trivial applications of common sense are indistinguishable from wisdom
From the article:
That might be the greatest value of these simulations. The impossibility of making truly accurate predictions suggests that large societies should be conservatively governed. Those of us who are interested in developing alternative societies should, in my opinion, start small, work slowly and hope to achieve something lasting over the course of generations. I discuss this sort of thing in my manifesto.
----------
Manifesto for the Peoples of the Third Millennium
Check out this cellular automaton which I made which makes some cool graphics:
www.geocities.com/enriqueeder/trip.html
Each pixel is a cell in the automaton. Each cell has 3 quantities each of which has a value between 0 and 255. The quantities correspond to the amount of red, green and blue in the color of the cell.
The color of each cell in the next frame of the simulation depends on its current color and the color of its neighbors in the current frame. The rule is that each quantity (red, green and blue) has an enemy or inhibitor quantity. For example green is by default the enemy of red, so the more green a cell's neighbors have in the current frame, the less red that cell will have in the next frame. Red is also the enemy of blue, and blue is the enemy of green. So each quantity has an enemy.
The simulation is seeded with a randomly colored cell by clicking on the black screen. To run the simulation, click the Go button. To stop it, click the Stop button. To advance just one frame click the Step button.
If you click the Design button, a window will pop up where you can modify the parameters of the calculation. The Neighbors amount determines how much the amount of the enemy quantity in a cell's neighbors affects that cell in the next frame. The Self amount determines how much the cell stays true to its current color. The Enemy amount affects how much one quantity is affected by its enemy quantity. The Direction button flips the quantities' enemies.
The unexpected result is trippy swirling patterns as red chases green, green chases blue and blue chases red.
I would like to see simulations of the slashdot community's overall response to moderation, membership fees, advertising, etc. Also, simulations of diverse markets of computer users in selecting operating systems would be interesting. Will answer questions like "will MS rule the world?", "are all the Linux companies doomed?", "is Steve Jobs insane?", and of course, "is BSD dying?"
Karma: Good (despite my invention of the Karma: sig)
If it were necessary to simulate the "unbirth" of a specific adult, that would imply that humans have enough uniqueness to make statistical models invalid.
sPh
sPh
The 'future' data could be physically withheld from the modeller, they could have been in a box since 1986. The pre 1986 data could have been collated, or simply stored, in 1986, to avoid being 'future' tainted. And wait for it .... they could even use 1986 computers :).
Be Free: Free Software Tuition
/.
A fatal flaw of this simulation (as a model of real society, that is) is that it includes the "Cincinatus" characters - the incorruptible agents - but does not include the "Dillingers" - agents who are not deterred by punishment, of themselves or of others.
I have found over the years that people who are not influenced by "common sense" (or even an informed sense of self-preservation) are much more common than incorruptible people. Luckily (perhaps) these people more commonly are obsessed with greed than killing, or we'd have a lot more mayhem and a few less rich people.
Thus, the simultation should include agents that are not influenced by the arrest rate, and the model will probably become cyclic instead of trending to a fixed equilibrium.
Your statement that "the simulation is accurate" is unfounded, as any serious study of real behaviour in a police state will show. The Chinese shoot homosexuals and drug addicts; yet they still occur just as frequently as in other nations with less draconian laws. The US is "soft on crime" according to the Immoral Minority, yet our crime rates continue to drop.
But of course, anyone who thinks humans are simple agents with simple motivations is very unobservant.
--Charlie
CS nerds might be in a good position to end the recession. We know how to do big simulations and distributed computing and how to mine for data to feed a simulation. We know how to run several simulations in parallel, each representing a different course of economic intervention.
The economy is driven primarily by human actions and decisions. In principle, humans could all agree that recessions are bad, and each tweak our behavior to end the damn thing. Given how much suffering the economy can cause, it seems ridiculous to leave it entirely to chance.
It may turn out that benign interventions are impossible because of conflicts of interest (an individual's own interests dictate behavior that prolongs the recession or injures society, what the economics folks call a tragedy of the commons). But it might at least merit investigation.
My own small effort in this direction appears in my sig.
WWJD for a Klondike Bar?
I wanted to point out to /.s out there that there is quite a bit of software available to explore artificial societies on your own. I also wanted to say that I have had the rare privledge of working w/ these folks for many years and all of the positive comments (and none of the negative ones :-)) in this thread are right on.
Anyway, Ascape, a software framework for agent-based modelling that I developed is available for download at the Brookings Website. Many other interesting models are also described there..
Its all in Java, and the source code for versions of many models is available. (For the person who was complaining that the results aren't reproducable, this will prove you completely wrong. In fact, Epstein and Axtell and others in the field have spent a lot of time thinking about how models can be independently understood and verified.)
Someone has allready mentiond JASSS; there is an article in that journal on Ascape.
The Ascape build on the Brookings website is now quite old. I joined the Bios Group some time ago and we've been improving and enhancing Ascape as part of our work in using complexity science in "real world" applications. So there should be a new public release RSN, but the version on the website now is relativly robust and has a lot of features. Note that the mailing list at the Brookings website appears to be down at the moment.
What other entertainment is there? I'm bored and want to be amused, without paying "them".
"I found some things in Epstein and Axtell that were not, or should not have been, replicable from the information given in their book."
More detail would help, I couldn't find any more information on your website. You may want to distinguish between a) not replicable and b) not replicable by you, potentially very different things. I reimplemented all of the chapter two and three models in Java and while I did have the advantage of having the authors next door, I found that I was able to achieve the same results quite easily without asking many questions. This situation seems to me to be very akin to any kind of research endeavor, where a paper describes just the critical issues. The implementations details (think lab aparatus, etc..) really cannot be usefully and completely covered in a general treatment. But the results are very reproducable, and I must say, quite robust.
It is interesting that you say that you are using a discrete event simulation environment in your work. Generally, these models have much more of a time-step quality, though even that doesn't really capture things completetly. Discrete event approaches have artifacts and issues of their own of course; I suspect that a lot of your issues are occuring along this boundary, but that's just a guess.
This story should have been from the hari seldon dept.
main(c,r){for(r=32;r;) printf(++c>31?c=!r--,"\n":c<r?" ":~c&r?" `":" #");}
Hmmm. this sounds vaguely farmilular...like an AI for such games as Age of Empires or Starcraft. Maybe there is someone clicking a mouse making me type this right not. In which case, they would not want me to find out the truth so they would try to kill me. Why oh why did I take the red pill.
I once shot a man who posted too many, "Imagine a beowulf cluster of these"
Its "the esteemed Brookings Institution" not "Institute."
;-)
Otherwise, your post is a little too trollish and ad hominem to bother responding to, sorry.
There is yet another variation over at:
EVLU
This one is for competing players, and features true strategy and populations that evolve.
..the distinction between explanation and prediction is especially important here.
That said, the possiblity of prediction on some level is certainly possible. For example, if you were to run the Artificial Anasazi model forward after calibrating it properly, the results could be just as predictive if not more so than a traditional model -- say a PDE model.
But the basic point stands out; precisely because complex adaptive systems are so path-dependent and involve so many loosly coupled pieces, actual prediction of future events ala Seldon is very questionable, unless you think about it in terms of this classic joke:
"I have a complete model of the world economic system. It's life-size and runs in real-time."
That is, you can begin to throw more and more things in to your model, but at some point you lose the explanatory power that the model gave you in the first place. I've seen this quite a bit; people throw the kitchen sink in and the model becomes a muddled mess. The real value of these models, especially at this point, is in their explanatory power; ie.e they help you to generate (Epstein's usage) phenomenon you observe in the real world with quite simple rules.
This in turn helps you to understand how relativly simple processes may be at play in complex phenomenon, and may even give you some ideas about how to work with them. In this way, these models could have a real policy impact in a much more engaged and robust way than the Club of Rome example offerred in another thread.
Depends on the model. In stock markets, we have self-defeating prophecies; if everybody knows a stock is going up, it's too late, it's already up. In other situations you may have self-fullfilling prophecies - 'There's going to be a war at some point, we know it, they know it, so we better attack first.'
The trick is knowing which is which...
The word "chaordic" is used as defined by Dee Hock (the person behind VISA) at http://www.chaordic.org and in his book "Birth of the Chaordic Age", which is essentially processes at the boundary between CHAos and ORDer and the social implications for how to design effective and responsive organizations for a dynamic society. The focus will be specially on computer simulations to support part of the goal defined here http://www.chaordic.org/who_hist.html#FourCond of: "Development of visual and physical models of chaordic organizations so that people have something to examine, experiment with, and compare to existing organizations. The models must contain the ethical and spiritual dimensions generally lacking in current models. In addition, computer simulations will need to be created to allow people to quickly see how clarity of purpose and principles allow institutions to self-organize, evolve over decades, and link in new patterns for an enduring constructive society."
People are invited to join the mailing list if they want at this page http://mail.freesoftware.fsf.org/mailman/listinfo/ simulchaord-discuss
if you want to contribute to project related discussions or submit
snippets of code (with the understanding contributions will be archived
and can be incorporated into the project under the GPL license). I have been posting some artificial life links there related to modelling social systems to get things started -- one of the first was a link to the Atlantic Monthly article discussed in this Slashdot thread. For now, I am using
use the list to record my own musings on related simulation issues
including design, architecture, and use cases. I will also be posting my experiences as I try to create such simulations. Feel free to lurk for a while or chime in.
Here is a page leading to the entire mailing list archives (aroudn twenty messages so far): http://mail.freesoftware.fsf.org/pipermail/simulch aord-discuss/
The main project page is here: http://savannah.gnu.org/projects/simulchaord/ Cooperative development of releases of code is hosted on Savannah using CVS although I haven't yet put up any content (files or homepage) besides what's archived in the mailing list.
At the moment I am looking at using Swarm http://www.swarm.org as the base -- although I may just use Python instead -- or even use both for different aspects.
A 21st century issue: the irony of technologies of abundance in the hands of those still thinking in terms of scarcity.
...CAs usually have simultaneous (parallel, synchronous) update rules, whereas agent-based models (asynchronous) typically don't.
To give a concrete example, in Conway's life each agent (Cell) determines each period wether it will be dead or alive in the next period and then those values are updated all at once. In contrast, in a model like sugarscape, an agent who "dies" would typically remove itself immediatly.
These kind of seemingly insignificant implementation details can have huge effects on model outcomes and dynamics. This also means, btw, that one person's "artifact" can be another's "feature" and vis. versa.
nuff said =)
When in doubt, parenthesize. At the very least it will let some poor schmuck bounce on the % key in vi. (Larry Wall)
This is simply an interesting, albeit apparently unintentional application of newtonian mechanics to social studies.
You know, the whole hoopla about how the behaviour of a system can be foretold if its initial state is known in entirety.
In theory, however, this does provide some interesting food for thought. If a contained system could be found, its initial state recorded, and permutations of its subsequent development run, we would have the first simulation of parallel universes.
Assuming, of course, that Newtonian mechanics allow for permutations. Which they don't. And assuming that it would be at all possible to record absolutely the initial conditions of a contained system. Impossible. As many of us know, the mere act of observation counts as interference in any experiment.
All the same, an interesting article if not revolutionary. Hey, it got me musing, didn't it.
Blearf. Blearf, I say.
sPh
The main problem with models like these is that they do not often take into account the dynamic nature of the "rules" that govern the simulated people. In the real world, people are able to change the rules that they live by, self-programming in a sense. For example, if we were to run a model that used the "rules" that governed race-relations in the 1800 and attempt to run that simulation forward to today, we would find that the end result is drastically different than the world we live in today, becuase the rules themlesves are evolving as the simulation moves forward. Maybe when simulating frog populations, this kind of rule-changing is less common, but when simulating people, it will always happen.
People have the ability to see the broader picture and alter the way the work in it. For example, in the scenario from the article where any particular square bases it's actions on the squares next to it, a "human" square would base it's rules on the squares next to it, BUT also on the makup of the board as a whole.
Once the simulators begin to allow the rules themselves to change, then we will see some really amazing results.
"Your superior intellect is no match for our puny weapons!"
Epstein and Axtell's book "Growing Artificial Societies," bears all of the typical marks of what we call bad science. While I am making no statments about Epstein and Axtell as human persons or as intellectuals, I do clearly state that their book is not good science.
The complaint about their lack of understanding of computer science issues was a throw-away comment and was irrelevent to the substance of my main beef with the book --- results that are not reproducable, and conclusions drawn from insufficient data. That is why I recommended that people not purchase Epstein and Axtell's book.
I wonder if simulations can deal with events like September 11? That has changed society in many countries.
Do they just count on things leveling out over time?
What about something extremely bad happening, like a large asteriod hitting the Earth? I think as long as nothing major happens, simulations could be accurate. But all it would take is aliens landing on the White House lawn or a widespread plague to totally throw off any mathmatical equation(s).
Since I don't code, I do thought experiments. Lately, they've been about "turnkey" operations large enough for thousands. I imagine thought experiments of this sort as as common as hemeroids. Anyone else into Douglas Hofstadter's "Alternative State of the Union"?
Keep the aspidistra flying!