I keep looking for something to see.. Since I am not a programmer, nor an economist etc., the figures seem alien in themselves, how exactly the agents got the results they did. What were the conditions? How come you think such an environment is comparable in any way with some human economic environments? |
Do I sound negative? Sorry.. I do not mean it in that way. I just think if such elements as the above were given, it would be easier for more people to take part, and criticise... |
[周易] |
Hi, Reading what you've said above, I'd say you have three specific points : # How do we get our results? # What are the background assumptions we're making? # How do we justify that our model is accurate? OK, briefly : *1) We get our results from computer simulation. That is, we have a computer program which represents information about a number of agents. Think of it as something like a video-game world, without the flashy graphics. In that world, the computer-controled agents work to gain resources. They consume resources. And in some of the worlds, they share their resources through one of several strategies. The SelfishForagers never share the resources they gain. The GiftAgents donate their spare resources to neighbours who need them. The BarterAgents barter spare resources for resources that they in turn need. And the MoneyAgents auction their spare. : There are no human players in this game world. The computer controled agents spend a number of rounds working, consuming, and sharing resources. Some agents will run short of the resources they need, and these agents will die. After a certain amount of time, we look at the result. We count how many of the agents are still alive. How much they worked to create. How much they consumed etc. When you look at a page with results such as GiftAndSelfishForaging, you can see the parameters for the simulation : how many agents we started with, how many time-steps the simulation was run for, etc. Then you see the performance of populations of the different agent-types. *2) There many background assumptions we make. The most important thing to remember is that we build models to test specific questions about the sharing strategies. And the models are as simple as we think we can get away with to test those questions. : Right now we aren't trying to produce a general model of the British or Brazilian or American economy. What we're trying to do is look at what differences, specific sharing strategies make when compared against each other. So the assumptions are very simplistic : ** All agents work to create resources. ** Different agents can have slightly different capacities to create each resource, and slightly different consumption rates. (Different enough that some agents are unviable and would die without external support.) ** There are no resources that require co-operation or co-ordination of a number of agents to produce. ** Agents all use the same strategy, and never try to cheat or game the system. (Though we will be moving towards economies with mixed or polymorphic strategies, where agents may cheat.) :Then there are the assumptions about how each strategy is played. The fine details are in the code, but we aim to document the strategies in pseudo-code so you don't need to be a programmer to understand. If you find any of these too sketchy, please ask. *3) The short answer is, we can't justify that our model is accurate in advance. Models are always abstractions. And may always be flawed because we've forgotten something. What we believe is that we've got some fairly intuitively plausible assumptions to get the conversation started. And people who disagree and want offer helpful criticism (see DialogueOfModels) have something to get their teeth into. There are specifics, which can be pointed to. And you can, for example, request that a particular assumption is changed. Hope this helps. regards PhilJones It seemst to me, that you are simulating the ancient economics of a hunter-gatherer-world, where the means of production are the people themselves. I do not understand, how money could be simulated in such a world. In my opinion the invention of money is not possible without the concept of property (for means of production). I suggest the following extensions to the model : means of production, like water resources, land, growing vegetables could be simulated as agents which produce specific but do not consume goods. means of production (including humans) may be owned by other agents. ownership may be partial (like a job-contract, ownership of the production of a specific good, where money or other goods will be paid for). ownership must be tradeable. Without simulationg property you will not be able to simulate the driving force behind work (WhyWork): artificial scarceness (there are growing enough apples for you and me, but unfortunately they all belong to me) In general I am very skeptical about the benefit of simulations of human culture, which, in my opinion, should be viewed at in a holistic way (where simulation is an atomistic view). Maybe there are enough real world arguments, see for example http://www.fastcompany.com/magazine/29/paranoia.html -- ThomasKalka Thanks for these points Thomas. (And welcome to Optimaes :-) Yep, this version of the simulation is very basic. I want to start from the ground up so I'm starting with basic hunter-gatherer type agents. I agree, it will certainly be interesting to add the means of production you're talking about. And I can see why we need to go there. But I'm curious why you say money is impossible or meaningless without them. I've never heard that, and it isn't immedietely intuitive to me why money shouldn't support division of labour in a world of hunter-gatherers. Could you say a little more about why you think this? As to the general skepticism of simulation, there are several responses. One is that it's what I do. It's something which speaks to me. And as far as my own rhetorical project is concerned. I think it will speak to people I want to speak to who haven't been touched by the other real world arguments. Personally I don't like the word "holistic" much. For me it can sound like an attempt to avoid analysis : "everything's interconnected and interdependant so you can't prove this is what's going on, but take my word for it". I do believe in the virtues of analysis, of breaking things down into their constituants and looking at the details. But rather than "atomism" I prefer to see this as systems thinking. I break things down into their constituent parts but I also care about their interactions and interdependencies, and about the overall statistical properties and behaviours of the group. For me, simulation is exactly a tool for trying to understand that complexity. I can start with atoms, and program the interactions, and observe the statistical / holistic properties. Nevertheless, this is a project that encourages dissent. So please feel free to push back and make an alternative case. Also, I've been writing a whole lot of stuff about Gift economies over on my personal wiki today which you might be interested in : http://www.nooranch.com/synaesmedia/wiki/wiki.cgi?GiftEconomy/Hidden -- PhilJones In a world of hunter and gatherer anyone is able to produce anything and there is no big difference in money to other goods, except the better transportability and endurance. In a world of property owners money is the one and only commodity the average agent need to survive. -- ThomasKalka OK. So I understand your argument like this : hunter-gatherers have no need for money because it's main virtue is portability? And as all their goods are portable anyway they could equally use barter? Isn't there also a question of co-ordination? : A wants something from B who wants something from C who wants something from A. Barter won't work because no pair want to exchange what they have for what the other has. But with money they can. (Having said this, I haven't managed to demonstrate this situation yet in Optimaes, so maybe you're right. :-) -- PhilJones |
Can't locate object method "startform" via package "CGI" at optimaes.cgi line 1479.
For help, please send mail to the webmaster ([no address given]), giving this error message and the time and date of the error.