Inferring interaction domains of collectively moving agents with varying radius of influence

Transfer entropy (TE) has proven to be an effective tool for determining the causal connection between two processes. For example, TE been used classify leader and follower agents in collective dynamics Vicsek model (VM). However, previous results have limited interaction radii, which are precisely same among all agents, is not realistic practice. Here, we propose a modified version of VM where domains agent can influenced by others vary from agent-to-agent, matches more closely with real-life setting physical traits. We demonstrate that vs cut-off technique robust efficacious maximum distance at interacting transfer information system. find different domains, derivative average inward determine individual agent’s domain. In system numerous shown predicting domain each was randomly selected Gaussian distribution.