Catálogo de publicaciones - libros

This paper develops a model for precisely defining how an agent interacts with objects in its environment through the use of its capabilities. Capabilities are recursively defined in terms of lower-level capabilities and actions, which represent atomic interactions with the environment. Actions are used to represent both sensors and effectors. The paper shows how the model can be used to represent both software and physical agents and their capabilities. The paper also shows how the model can be integrated into the Organization-based Multiagent Systems Engineering methodology.

Allocation of goal responsibilities to agent roles in Multi-Agent Systems (MAS) influence the degree to which these systems satisfy nonfunctional requirements. This paper proposes a systematic approach that starts from nonfunctional requirements identification and moves towards agent role definition guided by the degree of nonfunctional requirements satisfaction. The approach relies on goal-dependencies to allow potential MAS vulnerabilities to be studied. In contrast to related work where organizational patterns are imposed on MAS, roles are constructed first, allowing MAS organizational structures to emerge from role definitions.

A number of concerns in multi-agent system (MAS)design have a crosscutting impact on agent-oriented models. These concerns inherently affect several system agents and their internal modeling elements, such as actions and goals. Examples of crosscutting concerns in MAS design encompass both internal and systemic properties, such as learning, mobility, error handling, and security. Without an explicit modeling of such MAS properties, designers can not properly communicate and reason about them and their broadly-scoped effects. This paper pre sents a meta-modeling framework for supporting the modular representation of crosscutting concerns in agent-oriented design. The framework is centered on the notion of aspects to describe these concerns. It also defines new composition op-erators to enable the specification on how aspects affect the agent goals and ac-tions. The proposed framework is a result of our previous experience in both using aspect-oriented techniques for MAS design and implementation, and integrating aspect-oriented abstractions in an agent-oriented modeling language, called ANote.

This paper presents a proposal for modeling agent mobility with UML sequence diagrams. The notations used to model agent mobility are focused on capturing agent creation, mobility paths and current agent location. Four approaches are described and compared according to their clarity, the space needed for graphics and their expression of mobility. In a case study, the most suitable solution of the proposed notations for the given scenario is elaborated.

Governance means that specifications are enforced dynamically at application runtime. Governance framework is a technique to design and implement an extensible interaction specification for a family of open systems. This specification can be refined for particular applications. We based this proposal on object-oriented framework concepts and adapted them for distributed agents and interactions. A governance framework structures the extensions of open system instances as variations in interactions among agents, defined as templates. Templates are used to gather core implementation and extension points. Extension points are ”hooks” that will be customized to implement an instance of the governance framework. During framework instantiation, templates are refined to concrete interaction specification. As a proof of concept experiment, in this paper we propose a framework for instantiating supply chain management applications as open systems.

The vision of agents working together on the Internet, in virtual organizations, is one that is increasingly common. However, one of the issues is the regulation of the participating agents and their behaviour. A substantial body of work exists that investigates agent societies and agent organizations, including work on , such as and . However, although such work provides concrete tools for specifying and enacting institutions, there is a lack of clear documented guidance to designers who are using these tools. In this paper we describe a methodology for developing an institutional structure for multi agent systems. This methodology captures the knowledge and experience within the Islander group, and integrates it with the methodology.

This paper describes an agent-oriented requirements engineering approach that combines informal models with formal specifications in the multiagent system specification formalism CASL. This allows the requirements engineer to exploit the complementary features of the frameworks. can be used to model social dependencies between agents and how process design choices affect the agents’ goals. CASL can be used to model complex processes formally. We introduce an intermediate notation to support the mapping between models and CASL specifications. In the combined -CASL framework, agents’ goals and knowledge are represented as their mental states, which allows for the formal analysis and verification of, among other things, complex agent interactions where agents may have different goals and different (incomplete) knowledge. Our models can also serve as high-level specifications for multiagent systems.

Agents are becoming a popular technology for the development of distributed, heterogeneous and always available systems. The application of agent technologies requires extensions to the existing object-oriented modeling languages to accommodate agent-related abstractions such as roles, organizations and environments. If it is difficult to analyze and establish the well-formedness of a set of diagrams of a UML-like object-oriented modeling language, it gets far more complex when the language is extended to add a set of agency related abstractions. In order to tame such complexity, we propose an ontology-based method for analyzing MAS specifications described using a modeling language that extends UML to accommodate the agency characteristics. The method proposes a two-phase approach that covers different sets of MAS design properties. These properties are the ones related to each individual diagram and the ones associated with pairs of diagrams. The later take into consideration the interdependencies between diagrams. The method also provides features that allow the suggestion of some design guidelines which may improve the MAS design quality.

Recently, multi-agent systems have proved to be a suitable approach to the development of real-life information systems. In particular, they are used in the domain of safety critical systems where availability and reliability are crucial. For these systems, the ability to mitigate risk (e.g., failures, exceptional events) is very important. In this paper, we propose to incorporate risk concerns into the process of a multi-agent system design and describe the process of exploring and evaluating design alternatives based on risk-related metrics. We illustrate the proposed approach using an Air Traffic Management case study.

Complex and distributed nature of multi-agent systems (MASs) makes it almost impossible to identify of all requirements at the beginning of the development. Hence, development of such systems needs an iterative and incremental process to handle complexity and the continuously changing na-ture of the requirements. In this paper, a test driven multi-agent system devel-opment approach that naturally supports iterative and incremental MAS con-struction is proposed. Also a testing framework called as SUnit which supports the proposed approach by extending JUnit framework is introduced. This framework allows writing automated tests for agent behaviors and interactions between agents. The framework also includes the necessary mock agents to model the organizational aspects of the MAS.