Agent-based modeling

Agent-based modeling (ABM) is a computational approach where a system emerges from the interactions of many autonomous entities—called "agents"—each following simple rules. Rather than solving equations from the top down, ABM simulates from the bottom up, letting patterns emerge through individual behavior.

Each agent has properties, makes decisions, and responds to its environment and neighboring agents. Collectively, their actions produce temporal dynamics that often surprise their creators: traffic jams from lane changes, market crashes from individual trading, or disease spread from contact patterns.

ABM thrives in domains where central control doesn't exist or where networks of actors matter more than aggregate statistics. Bacteria competing for resources, firms in markets, voters in elections, even cells in biological tissues—all are natural ABM subjects.

The power lies in bridging human behavior, computational intelligence, and real-world complexity. Rather than assuming rational actors or equilibrium, ABM can capture bounded rationality, incomplete information, and feedback loops that traditional models miss.

Tools like NetLogo, Repast, and Mesa make ABM accessible. Results are often visualized as animations or interactive dashboards, grounding abstract rules in visible spatial patterns.

Related

Simulation, Complex systems, Emergence, System dynamics, Monte Carlo method, Behavioral economics