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…

This is a summary of Swarm Intelligence: A Whole New Way to Think About Business by Eric Bonabeau and Christopher Meyer, from the Harvard Business Review May 2001.

Eric Bonabeau begins his article, “Swarm Intelligence: A Whole New Way to Think About Business” by introducing the subject and going over a few simple path-finding and foraging examples. These include both business, airlines, and insects, ants and termites. He initially emphasizes the simple rules and complete lack of managerial organization. One might mistake Bonabeau for a communist. He continues by showing example after example where traditional managed solutions were unable to adapt and find an optimal setup and a swarm intelligent answer did. While one could argue that the worker is emphasized over the manager, and one may wonder what the role of a manager is in a business structure modeled after swarm intelligence, the ideas that all employees should and can be able to come up with new ideas and implement them, due to their innate qualities and peer support, is very capitalistic.

Focusing on three advantages of swarm intelligence, flexibility, robustness, and self-organization, Bonabeau admits that there can be many forms of swarm intelligence and some work better than others for certain situations. The examples he gives are how ants forage for food. One of three techniques are used: mass, paired, and group foraging. Depending on the size of the colony, some of these are more efficient, that is more flexible and robust. So the question is not only “should swarm intelligence be applied to this problem?” but also “how should swarm intelligence be applied?”