Multiagent Systems

What is a multiagent system?

A multiagent systems is a collection of automonous agents acting independently using only local information and with, possibly, the ability to communicate with each other. That is, they are people in the world, or bees in a hive, or cells in your body, or neurons in your brain, or corporations in the market. We are intereset in engineering multiagent systems whose emergent behavior satisfies some pre-defined properties. In this way, we are not much different from an Economist who tries to set the rules of commerce so as to create the most efficient allocation of resources. The main different is that our systems are composed of both human and artificial agents, and we can program the artificial agents to behave exactly as we wish.

As such, researchers in multiagent systems blend results from game theory, economics, and complex systems along with the tools provided by artificial intelligence and distributed programming algorithms, helped by a good dose of semantic web terminology when building software systems, in order to design efficient multiagent systems.

You can also download my book Fundamentals of Multiagent Systems and get a solid introduction to the field. If you are interested in the web-services side of things I recommend Service-Oriented Computing: Semantics, Processes, Agents.

How to I submit a story to this website?

Read the about page.

Is there another way to search over the links on this site?

You can also view my del.icio.us bookmarks but note that those do not include the longer blog entries from this site. On the other hand, they have a more sophisticated search language and you can also see how many other people have bookmarked each page.

Who runs this?

Jose M Vidal, check out my lab's blog.

I have another question for the FAQ!

Then, add it to the comments below. I will try to answer it.