Introduction to Adaptive Systems

Intro to Adaptive Systems (aka “Introduction to Evolutionary Dynamical Systems”) centers on the construction and use of agent-based adaptive models study phenomena which are prototypical in the social, biological and decision sciences. These models are "agent-based" or "bottom-up" in that the structure placed at the le vel of the individuals as basic components; they are "adaptive" in that individuals often adapt to their environment through evolution or learning. The goal of these models is to understand how the structure at the individual or micro level leads to emergent behavior at the macro or aggregate level. Often the individuals are grouped into subpopulations or interesting hierarchies, and the researcher may want to understand how the structure of development of these populations affects macroscopic outcomes.

The course will start with classical differential equation and game theory approaches. It will then focus on the theory and application of particular models of adaptive systems such as models of neural systems, genetic algorithms, classifier system and cellular automata. Time permitting, we will discuss more recent developments such as sugarscape and echo.