Catálogo de publicaciones - libros

The rational agent community uses Michael Bratman’s as its theoretical foundation for the development of its agent-oriented BDI languages. We present an alternative framework based on situated action and activity theory, combining both BDI and activity-based modeling, to provide a more general agent framework. We describe an activity-based, as opposed to a goal-based, BDI language and agent architecture, and provide an example that shows the flexibility of this language compared to a goal-based language.

The construction of multi-agent systems is starting to become a main issue in agent research. Using the Computer Science point of view, the development of agent systems has been considered mainly a problem of elaborating theories, constructing programming languages implementing them, or formally defining agent architectures. This effort has allowed important advances, including a growing independence of Artificial Intelligence. The results have the potential to become a new paradigm, the agent paradigm. However, the acceptance of this paradigm requires its application in real industrial developments. This paper uses this need of addressing real developments to justify the use of software engineering as driving force in agent research. The paper argues that only by means of software engineering, a complex development can be completed successfully.

This paper surveys the state-of-the-art in developing multi-agent systems, and sets out to answer the questions: “what are the key current issues in developing multi-agent systems?” and “what should we, as a research community, be paying particular attention to, over the next few years?”. Based on our characterisation of the current state-of-the-art in developing MAS, we identify three key areas for future work: techniques for integrating design and code; extending agent-oriented programming languages to cover certain aspects that are currently weak or missing (e.g., social concepts, and modelling the environment); and development of debugging and verification techniques, with a particular focus on using model checking also in testing and debugging, and applying model checking to design artefacts.

A multiagent system (MAS) structures a software system as a set of autonomous agents that interact through a shared environment. Software architecture is generally considered as the structures of a system which comprise software elements and the relationships among the elements. So there is a clear connection between MAS and software architecture. In our research, we study situated MAS, i.e. systems in which agents have an explicit position in the environment. We apply situated MAS to domains that are characterized by highly dynamic operating conditions and an inherent distribution of resources. We use an architecture-centric approach for developing such MAS. From our experiences with building various applications, we have developed a reference architecture for situated MAS. The reference architecture provides an asset base architects can draw from when developing new systems that share the common base of the reference architecture. In this paper, we explain our perspective on architecture-centric software development of MAS. We give an overview of the reference architecture and we show an excerpt of the software architecture of an industrial application in which we have used the reference architecture. The reference architecture shows how knowledge and experience with MAS can be documented and matured in a form that has proven its value in mainstream software engineering. We believe that this integration is a key to industrial adoption of MAS.

In the last years, social and organizational aspects of agency have become a major issue in multi-agent systems’ research. Recent applications of MAS enforce the need of using such aspects in order to ensure some social order within these systems. However, there is still a lack of comprehensive views of the diverse concepts, models and approaches related to agents’ organizations. Moreover, most designers have doubts about how to put these concepts in practice, i.e., how to program them. In this paper we focus on and discuss about the literature on formal, top-down and pre-existent organizations by stressing the different aspects that may be considered to program them. Finally, we present some challenges for future research considering particularly the openness feature of those agents’ organizations.

refers to an abstract and intangible concept. In modelling, intangible concepts can be embodied and made explicit. This allows to manipulate the abstractions and to build predictable designs. Business processes in organisations are in fact reducible to interactions, especially when modelling methods are employed. Business processes represented as interaction structures can appear at different levels of abstraction. There is a compositional coupling between these levels, and this necessitates a method that allows dynamic de/re-composition of hierarchically organised interactions. We introduce the novel concepts that allow interaction-based diagramming and explain the syntax and semantics of these constructs. Finally, we argue that a business process composition with interactions allows more organisational flexibility and agent autonomy, providing a better approach in complex and dynamic situations than current solutions.

Design patterns are templates of general solutions to commonly-occurring problems in the analysis and design of software systems. In mature development processes, engineers use and combine these patterns to work out those parts of their systems that correspond to well-identified issues in their domains. The design of new structures is just concerned with those aspects that are specific for their projects and with the glue between different components. Model driven development approaches can benefit of design patterns to improve the building of models and their transformations; at the same time, design patterns can take advantage in this kind of approaches of a better integration in the overall development process. In the case of Agent-Oriented Software Engineering, design solutions for agents and multi-agent systems have been also described in the literature. However, their application and transformation to code largely relies on manual processes. This paper proposes a framework that includes repositories of patterns that can be reused in different projects and processes to generate models and code for multi-agent systems on different target platforms. Instead of focusing on low-level issues, our approach positions the abstraction level of these design patterns at the intentional and social features that characterize multi-agent systems. The paper illustrates this framework with a case study about the development of the models of an agent-based system for collaborative filtering of information.

Ad hoc networks are a type of computational system whose members may fail to, or choose not to, comply with the laws governing their behaviour. We are investigating to what extent ad hoc networks can usefully be described in terms of permissions, obligations and other more complex normative relations, based on our previous work on modelling norm-governed multi-agent systems. We propose to employ our existing framework for the specification of the laws governing ad hoc networks. Moreover, we discuss a software infrastructure that executes such specifications for the benefit of ad hoc network members, informing them of their normative relations. We have been developing a sample node architecture as a basis for norm-governed ad hoc network simulations. Nodes based on this architecture consider the network’s laws in their decision-making, and can be individually configured to exhibit distinct behaviour. We present run-time configurations of norm-governed ad hoc networks and indicate design choices that need to be made in order to fully realise such networks.

The research on exception handling in Multi-Agent Systems has produced some advanced models to deal with ‘exceptional situations’. The expression ‘agent exception’ is however unclear across the literature, as it sometimes refers to extensions of traditional exception models in programming languages, and sometimes to organizational management mechanisms with distinct semantics. In this paper, we propose a definition of ‘agent exception’ to clarify the notion and justify that specific research is necessary on this theme. We detail properties of this definition, revisit the traditional vocabulary related to exception in software design, propose an adequate agent architecture, and identify some research issues. This work is aimed at federating the endeavors on the question of exception management for Agent-Oriented Computing.

In this chapter we explore the role of regulation in joint activity that is conducted among people and how understanding this better can enhance the efforts of researchers seek ing to develop effective means to coordinate the performance of consequential work within mixed teams of humans, agents, and robots. Our analysis reveals challenges to the quality of human-machine mutual understanding; these in turn set upper bounds on the degree of sophistication of human-automation joint activity that can be supported today and point to key areas for further research. These include development of an ontology of regulatory systems that can be utilized within human-agent-robotic teamwork to help with mutual understanding and complex coordination.