Conversations are sequences of messages exchanged among interacting agents. For conversations to be meaningful, agents ought to follow commonly known specifications limiting the types of messages that can be exchanged at any point in the conversation. These specifications are usually implemented using conversation policies (which are rules of inference) or conversation protocols (which are predefined conversation templates). In this article we present a semantic model for specifying conversations using conversation policies. This model is based on the principles that the negotiation and uptake of shared social commitments entail the adoption of obligations to action, which indicate the actions that agents have agreed to perform. In the same way, obligations are retracted based on the negotiation to discharge their corresponding shared social commitments. Based on these principles, conversations are specified as interaction specifications that model the ideal sequencing of agent participations negotiating the execution of actions in a joint activity. These specifications not only specify the adoption and discharge of shared commitments and obligations during an activity, but also indicate the commitments and obligations that are required (as preconditions) or that outlive a joint activity (as postconditions). We model the Contract Net Protocol as an example of the specification of conversations in a joint activity.

No abstract

No abstract

Abstract. Building conversation protocols has traditionally been an art more than a science, as their construction is often guided by designers' intuition rather than by a principled approach. In this paper we present a model for building conversation protocols using inference principles that allow the computational specification and verification of message sequencing and turn-taking. This model, which is based on the negotiation of social commitments, results in highly flexible protocols that support agent heterogeneity while abiding by software engineering practices. We exemplify the specification of protocols using the contract net protocol, a common interaction protocol from the multiagent literature.

No abstract

In this paper, we describe several interesting design decisions we have taken (with respect to inter-agent messaging) in the reengineered CASA architecture for agent communication and services. CASA is a new architecture designed from the ground up; it is influenced by the major agent architectures such as FIPA, CORBA, and KQML but is intended to be independent (which doesn't imply incompatible). The primary goals are flexibility, extendibility, simplicity, and ease of use. The lessons learned in the earlier implementation have fed the current design of the system. Among the most interesting of the design issues are the use of performatives that form a type lattice, which allows for observers, who do not necessarily understand all the performatives, to nonetheless understand a conversation at an appropriate semantic level. Furthermore, we found it difficult to arrange a portion of the FIPA performatives in a lattice without duplication and complexity. We solved this problem by diverging from FIPA performatives by separating a lot of the performatives into a separate lattice of acts, which work with the performatives, but greatly simplify and enhance the semantic interpretation of the messages. Yet another innovation is the addition of several new fields within a KQML-style message header that allow for further semantic interpretation of the message by an observer who does not necessarily understand the content language. The traditional FIPA fields to, from, sender, and receiver have been extended to include agent (the requester of a service), and actor (the party designated responsible for performing the action specified in the performative/act). These new design considerations add a great deal of flexibility and integrity to an agent communications architecture.

In this paper, we arrange FIPA's ACL performatives to form a subsumption lattice (ontology) and apply a theory of social commitments to achieve a simplified and observable model of agent behaviour. Using this model, it is straight forward to model agents' social commitments (obligations) based solely on observation of messages passed between the agents (such observation is supported by our agent infrastructure system). Furthermore, owing to the performatives being in a subsumption lattice, it is relatively easy for an observer to infer social commitment relationships even if the observer does not understand the details of messages or even the exact performatives used (so long as the observer has access to the performatives ontology).Although social commitment modelling of agent behaviour does not demand that agents are implemented using the social commitment model, our model nonetheless can be used in agent implementation to simplify the specification of agent behaviour through the use of "commitment operators" attached to the performatives (as policies) in the subsumption lattice.The primary contribution of this paper is to show how FIPA's performatives can be mapped onto a theory of social commitment to allow observable social behaviour and conformance to social norms.

No abstract