Influence-Based Autonomy Levels in Agent Decision-Making

摘要

Autonomy is a crucial and powerful feature of agents and it is the subject of much research in the agent field. Controlling the autonomy of agents is a way to coordinate the behavior of groups of agents. Our approach is to look at it as a design problem for agents. We analyze the autonomy of an agent as a gradual property that is related to the degree of intervention of other agents in the decision process. We define different levels of autonomy in terms of inter-agent influences and we present a BDI-based deliberation process in which different levels of autonomy can be implemented.