Modeling Communicative Behavior Using Permissions and Obligations

Abstract

In order to provide flexible control over agent communication, we propose an integrated approach that involves using positive and negative permissions and obligations to describe both conversation specifications and policies. Conversation specifications are described in terms of the speech acts that an agent can/cannot/must/must not perform based on the sequence of messages received and sent. On the other hand, conversation policies restrict how the specifications are used and are defined over the attributes of the sender, receiver, message content, and context in general. Other policies like management, social, privacy etc. are defined at a higher level of abstraction and restrict the general behavior of agents. Whenever they deal with communication, the higher level policies are translated into conversation policies using the syntax and semantics of the specific communication language being used. Agents use a policy engine for reasoning over conversation specifications and applicable policies in order to decide what communicative act to perform next. Our work is different from existing research in communication policies because it is not tightly coupled to any domain information such as mental states of agents or specific communicative acts.The main contributions of this work include (i) an extensible framework that can support varied domain knowledge and different agent communication languages, and (ii) the declarative representation of conversation specifications and policies in terms of permitted and obligated speech acts.