Catálogo de publicaciones - libros

Agents in an open system interact with each other based on (typically, published) protocols. An agent may, however, deviate from the protocol because of its internal policies. Such deviations pose certain challenges: (1) the agent might no longer be conformant with the protocol—how do we determine if the agent is conformant? (2) the agent may no longer be able to interoperate with other agents—how do we determine if two agents are interoperable? (3) the agent may not be able to produce some protocol computations; in other words, it may not cover the protocol—how we determine if an agent covers a protocol?

We formalize the notions of conformance, coverage and interoperability. A distinctive feature of our formalization is that the three are orthogonal to each other. Conformance and coverage are based on the semantics of runs (a run being a sequence of states), whereas interoperability among agents is based upon the traditional idea of . We present a number of examples to comprehensively illustrate the orthogonality of conformance, coverage, and interoperability.

Compliance is a property of an agent’s execution whereas conformance is a property of the agent’s design. In order to produce only compliant executions, first and foremost the agent must be conformant; second, it must also be able to interoperate with other agents.

Ontologies play a key role in agent communication and the emerging Semantic Web to define a vocabulary of concepts and their relationships. Different agents and web services may use vocabularies from different ontologies to describe their data. The current research on ontology mapping and ontology translation mainly focuses on how to map and translate vocabularies and associated data instances from one ontology to another. However, more complicated true statements, such as axioms (rules), are used or being developed to describe the relationships among the concepts. When extending one ontology using complicated true statements (theory) from another, we must confront the problem of , which is difficult because of the . In this paper, using an inferential approach we call , we show how to translate complex axioms between different time ontologies. We also prove the validity of our algorithm.

This paper presents complexity results for model checking formulae of (a logic to reason about time and knowledge in multi-agent systems) in concurrent programs. We apply these results to evaluate the complexity of verifying programs of two model checkers for multi-agent systems: and Verics.

This paper is concerned with the problem of obtaining predictable interactions between groups of agents in open environments when individual agents do not expose their logic. The most popular approaches to this in practise have been to model interaction protocols and to model the deontic constraints imposed by individual agents. Both of these approaches are appropriate and necessary but their combination creates the practical problem of ensuring that interaction protocols are meshed with agents that possess compatible deontic constraints. This is essentially an issue of property checking dynamically at run-time. We show how model checking can be applied to this problem.

The AgentSpeak agent-oriented programming language has recently been extended with various new features, such as speech-act based communication, internal belief additions, and support for reasoning with ontological knowledge, which imply the need for belief revision within an AgentSpeak agent. In this paper, we show how a polynomial-time belief-revision algorithm can be incorporated into the AgentSpeak interpreter by making use of ’s language constructs and customisation features. This is one of the first attempts to include automatic belief revision within an interpreter for a practical agent programming language.

In this paper we propose a foundational ontology of the social concepts of organization and role which structure institutions. We identify which axioms model social concepts like organization and roles and which properties distinguish them from other categories like objects and agents: the organizational structure of institutions and the relation between roles and organizations. All social concepts depend on descriptions defining them, which are collectively accepted, and the descriptions defining the components of organizations, including roles, are included in the description of the organizations they belong to. Thus, the relational dependence of roles means that they are defined in the organizations they belong to.

This paper discusses the problem of in dynamic environments and critical situations. When a number of (distributed) agents have only partial access to information, the explanation(s) and conclusion(s) they can draw from their observations are inevitably uncertain. In this context, the efficient propagation of information is concerned with two interrelated aspects: spreading the information as quickly as possible, and refining the hypotheses at the same time. We describe a formal framework designed to investigate this class of problem, and we report on preliminary results and experiments using the described theory.

In [15,19] we showed how to combine propositional multimodal logics using Gabbay’s methodology. In this paper we extend the above mentioned works by providing a tableau-based proof technique for the combined/ fibred logics. To achieve this end we first make a comparison between two types of tableau proof systems, ( & ), with the help of a scenario (The Friend’s Puzzle). Having done that we show how to uniformly construct a tableau calculus for the combined logic using Governatori’s labelled tableau system . We conclude with a discussion on ’s features.

AgentSpeak is a well-known language for programming intelligent agents which captures the key features of reactive planning systems in a simple framework with an elegant formal semantics. However, the original language is too abstract to be used as a programming language for developing multi-agent system. In this paper, we address one of the features that are essential for a pragmatical agent programming language. We show how certain of AgentSpeak plans can be used to define various types of declarative goals. In order to do so, we first define informally how plan failure is handled in the extended version of AgentSpeak available in , a Java-based interpreter; we also define special (internal) actions used for dropping intentions. We then present a number of which correspond to elaborate forms of declarative goals. Finally, we give examples of the use of such types of declarative goals and describe how they are implemented in .

In this paper, we describe the declarative agent programming language Jadl (JIAC Agent Description Language). Based on three-valued logic, it incorporates ontologies, FIPA-based speech acts, a (procedural) scripting part for (complex) actions, and allows to define protocols and service based communication. Rather than relying on a library of plans, the framework implementing jadl allows agents to plan from first principles. We also describe the framework and some applications that have been implemented.