Agent-Based Simulation Overview

This one-day seminar provides a comprehensive introduction to agent-based simulation (ABS), which is currently the “hottest” topic in simulation modeling. In an ABS autonomous agents (people, vehicles, organizations, etc.), which have attributes and potentially complex behaviors, interact with each other and their environment over time toward the accomplishment of their goals. This allows an agent’s behavior to depend on the current and past states of its environment, rather than being “scripted,” which permits much more complex behaviors to be represented as compared to traditional models.

ABS has been successfully applied to a diverse set of problems, and improved software packages have facilitated the model-development process. However, learning ABS on one’s own is difficult, at best, due to the genuine lack of specificity in the literature. Much of the confusion is due to the “smokescreen” of attributes that are often associated with ABS, including autonomy, sensing and interacting with the environment (including other agents), learning, adaptation, behavior based only on “local” information, emergence, and complex adaptive systems. We address this conundrum by discussing what we believe to be the real essence of ABS. The development of this seminar has benefited from funding by the U.S. Army.