Guidance and Law Policies in Multiagent Systems

摘要

Policies have traditionally been a way to specify properties of a system. In this paper, we show how policies can be applied to Organization- based Multiagent Systems Engineering (O-MaSE), specifically, in the OMACS meta-model. In OMACS, policies may constrain assignments of agents to roles, the structure of the goal model for the organization, or how an agent may play a particular role. We also show how traditional policies can be characterized as law policies. Law policies must always be followed by a system. Because of this inflexibility, law policies may constrain a multiagent system too much. In order to preserve flexibility of the system, while still being able to guide the system into preferring certain behaviors, we introduce the concept of guidance policies. These so-called guidance policies need not always be followed. When the system cannot continue with the guidance policies, they may be suspended. We show how this can increase performance while not decreasing flexibility of the system to adapt. Guidance policies are formally defined and, since multiple guidance policies can introduce conflicts, a strategy for resolving conflicts is given.