Agent-Based Models

Agents

An agent is a goal-driven "creature" that inhabits a constrained world.

Examples of agents include animals, plants, robots, and characters in a computer game.

Mobile agents can move about the world, while stationary agents must stay put. For example, in a market world vendor agents are stationary while shopper agents are mobile.

Agents interact with other agents. Examples of interactions include fighting, mating, and bargaining.

Agents typically make decisions based on limited knowledge of their world. For example, an agent might make false conclusions about the world based on observations of his village.

While the behavior of individual agents can be simple and boring (e.g., do what your neighbors do), the behavior of their society (i.e., the collection of all agents) can be complex and interesting. This is called emergent behavior.

Models

An agent-based model is a simulation of a world containing agents.

Each agent in the model encapsulates attributes and can execute procedures.

Note: According to this definition agents in a model are identical to objects in an object-oriented program. The only difference is that agents are objects that represent creatures as described above.

For example, and agent might have energy and position attributes. Asking such an object to execute a move procedure might change the values of these attributes.