Complexity


is it complex or just complicated?
an overlooked distinction

canonical examples
finding structure in disorder

multi-agent systems

dynamic and adaptive systems

emergent behavior
more than the sum of its parts










next





Complicated things are often confused with Complex things, but there is a subtle distinction:
Complicated -- comprises many parts
Complex -- involves many interactions
Which is not to say that one cannot also be the other. Also it should be noted that complex systems are NOT random systems. They may display chaotic behaviors but there is an underlying structure to be found.

The Canonical Examples section shows some "simple" complex systems, the Logistic Equation and the Lorenz Attractor. They illustrate the ideas that Complexity lurks in what seem to be very straight-forward models and where, while being completely deterministic, the end results are often so dependent on the starting state that they are not predictable in advance (sensitive dependence on initial conditions).

One area of great interest is that of Multi-Agent Systems. These are environments containing many (more or less) independent individuals that interact with one another, often in very simple ways. When viewed on a grander scale these systems can
sometimes exhibit aggregate behaviors that are very surprising. Complex systems can arise as the result of many simple interactions.

All of this leads us to the idea of Emergent Behavior. This can be anything from finding a new metric for a system quality that doesn't actually exist within the underlying components, to observing behaviors that are more than the sum of their parts. In many cases behaviors emerge solely from low level interactions and disappear when the system is reduced to it's component parts. For these systems the old scientific reductionist paradigm breaks down.