llor.nu is a giant game of Monopoly. The author does not come out and say this on the about page, probably for legal reasons, but the concept is pretty much the same, except it is done on a much larger scale (around the world) and you play one turn a day.

I don’t think the gameplay will be very interesting, I don’t even know if you make any decisions, but the concept of making games like this that are web based intruige me. I played AstroWars for a few months and thoroughly enjoyed it. You log on a few times a day, spend 5 minutes choosing what to spend your accumulated resources on, what to research, and where to send your ships. That’s all! It is imposssible to spend more than 5 minutes at a time, unless you analyze everything. It is the perfect game to get addicted to, because it can’t suck hours out at a time! The real joy comes when you finally build up a fleet large enough to squash your neighbor. This also happens at its own pace as it literally takes days for your ships to travel to another planet. Alliances quickly form and those left alone are soon destroyed. The game resets every other month or so, once someone reaches X amount of points, or an alliance reaches X amount of average points.

I am on the lookout for simpler ideas to implement. Combining my long time desire to design games, and my current passion for web development would be awesome. Who knows, maybe it could implement some sort of swarm intelligence?

As I think about swarm intelligence, I keep trying to figure out what it can be applied to beyond ant foraging and optimizing shipping routes. This idea just came to me.

Classical free market economists believe that a policy of hands-off or laize-faire is the best policy. Governments should not get involved with the economy, and over time, it will balance itself out and correct for any mishaps. Keynesian economists believe that it is appropriate and helpful for the government to act as a shock absorber to smooth out the economy’s climbs and dives. I think it may be possible to model the classical view using swarm intelligence.

Based on the presmise that classical ecnomics defies a centrilized unit of direction, most likely the government, this means that each player involved must make his own decisions. His objective is to optimize his own profits. He consults no other players and does not receive instruction from anyone. When every player in the market follows this “rule,” then following classical economics, the market should optimize itself.

Therefore it should be possible to come up with a set of rules that each player must follow, and given a large enough set of players, and a fairly realistic set of resources, a stable economy should arise.

Resources: These should be a fairly large set of varying, scarce resources. The demand for the products of every resource should outweight the supply. It is unimportant to fit resources into products such as “cars, books, food” but it may be useful to classify them as “normal” and “luxury” goods. This would then require that consumers have different tastes.

Every player, or firm, in the open market would be responsible for acquiring resources and converting them into products. Every product would have its own supply and demand curve. Players would have to decide what markets to enter and exit. Entering and exiting markets would depend on the profits to be had in each one.

After a fairly short period of time I believe the market would stabilize. However, what will happen when the availability of resources change? Or the demand for a specific product goes up? Or a new product is available? The previous players must reevaluate their current position and decide if it is worthwhile to switch markets. Hopefully, the swarm intelligence will find a way to optimize itself given any changing variables.

Judging the economy: in order for this to mean anything, one must be able to grade or rate an economy. How does one tell if an economy is doing well or poorly? the only thing I can come up with currently is to record some sort of GDP of the market. This should be good enough for a simple test.

I order to really compare classical economics with Keynesian economicsm, a similar test would have to be made with government influence. This would obviously not be based on swarm intelligence, and is therefore irrelevant to this discussions. This is fun is it not?

As you all know, because you read my last post, I am taking Swarm Intelligence this quarter. Here is the course description and definition.

Swarm Intelligence is a fairly new field in Artificial Intelligence. Swarm intelligence investigates how to obtain complex behavior from the cooperation of individual agents with relatively simple behavior. A popular example is ant behavior. A colony of ants exhibits behavior that is more complex than that of an individual ant. While individual ants exhibit extremely simply behavior, the ant colony as a whole solves shortest path problems, a hard problem in CS.

This is a research course. The majority of the course work will be concerned with reading research papers and scientific books from CS, Biology, Physics, and other areas which study systems which fall under the general category of swarm intelligence or complex systems.

There will be a course project in which groups of students study a system of their choosing. This system will be analyzed and formalized with the tools of swarm intelligence. A prototype will be produced which implements the behavior formalized.

I will be writing a series of summaries, responses, and papers on the topic and will be keeping a journal of them. This place seems as good as any, and you will all be able to read them. Is that not exciting? And so, here is the first summary…