Agents, pheromones, and mean-field models

Some agent-based models use digital analogs of insect pheromones for coordination. We clarify the position of this family of techniques in the spectrum of modeling tools. Inspired by statistical physics, we consider the nature of “mean-field models.” We show by analysis and simulation that pheromone models are intermediate between classical agent-based models and mean-field models. This position is not fixed, but can be adjusted by pheromone parameters (notably, the propagation factor). We discuss implications for the design of agent-based models.