Dynamic patterns formed by heterogeneous boid model composed of agent groups moving reversely

In this study, a heterogeneous boid model comprising many agents, which are divided into certain types, is examined. By varying the sensor area of the agent and the interaction within and between groups of agent types, the heterogeneous boid model generates various patterns. In this study, the patterns are tuned as type 1 and 2 agents gather and move reversely to each other. The proposed model is called as the retrograde model. This model generates an oscillatory pattern, where groups of different agent types oscillate reversely within agent clusters. As oscillation strengthens, the retrograde model forms fragile clusters with repeated fusions and fragmentations. To unify the agents into a stable cluster, the retrograde model is extended by introducing an additional type 3 agent that connects the existing type 1 and 2 agents. When tuned well, the extended retrograde model generates dynamic patterns with a unified long-lasting agent cluster in a wide range of agents.