The application of domain of danger in autonomous agent team and its effect on exploration efficiency

Prey animals in a group constantly face the trade-off between foraging gain and predation risk. The domain of danger concept was developed by ethologists to measure the predation risk taken by an individual within a group. In this work we make the analogy between information gathering for an autonomous agent team and food gathering for a group of animals to explore bio-inspired strategies for autonomous agent teams in an information gathering scenario. The sum of areas of domains of danger of the team is used as the measurement of risk taken by the whole team. Several movement rules that let an agent gather food nodes and/or shrink its domain of danger according to the current layout of the team and food nodes are proposed. The performance on foraging efficiency and team domain of danger of these movement rules are investigated by simulation. The results show that trying to shrink the domain of danger of each individual while no food can be gained in the near future doesn´t degrade the foraging performance of the team and can effectively lower the risk taken by the whole team by prevention of over-spreading.