Systems composed of interacting autonomous agents offer a promising software engineering approach for developing applications in complex domains. However, this multiagent system paradigm introduces a number of new abstractions and design/development issues when compared with more traditional approaches to software development. Accordingly, new analysis and design methodologies, as well as new tools, are needed to effectively engineer such systems.

The Grid and agent communities both develop concepts and mechanisms for open distributed systems, albeit from different perspectives. The Grid community has historically focused on "brawn": infrastructure, tools, and applications for reliable and secure resource sharing within dynamic and geographically distributed virtual organizations. In contrast, the agents community has focused on "brain": autonomous problem solvers that can act flexibly in uncertain and dynamic environments. Yet as the scale and ambition of both Grid and agent deployments increase, we see a convergence of interests, with agent systems requiring robust infrastructure and Grid systems requiring autonomous, flexible behaviors. Motivated by this convergence of interests, we review the current state of the art in both areas, review the challenges that concern the two communities, and propose research and technology development activities that can allow for mutually supportive efforts.

Trust and reputation are central to effective interactions in open multi-agent systems in which agents, that are owned by a variety of stakeholders, can enter and leave the system at any time. This openness means existing trust and reputation models cannot readily be used. To this end, we present FIRE, a trust and reputation model that integrates a number of information sources to produce a comprehensive assessment of an agent's likely performance. Specifically, FIRE incorporates interaction trust, role-based trust, witness reputation, and certified reputation to provide a trust metric in most circumstances. FIRE is empirically benchmarked and is shown to help agents effectively select appropriate interaction partners.

Abstract—This paper surveys and analyzes the state of the art of agent-mediated electronic commerce (e-commerce), concentrating particularly on the business-to-consumer (B2C) and business-to-business (B2B) aspects. From the consumer buying behavior perspective, agents are being used in the following activities: need identification, product brokering, buyer coalition formation, merchant brokering, and negotiation. The roles of agents in B2B e-commerce are discussed through the business-to-business transaction model that identifies agents as being employed in partnership formation, brokering, and negotiation. Having identified the roles for agents in B2C and B2B e-commerce, some of the key underpinning technologies of this vision are highlighted. Finally, we conclude by discussing the future directions and potential impediments to the wide-scale adoption of agent-mediated e-commerce. Index Terms—Agent-mediated electronic commerce, intelligent agents. 1

An important concept for intelligent agent systems is goals. Goals have two aspects: declarative (a description of the state sought), and procedural (a set of plans for achieving the goal). A declarative view of goals is necessary in order to reason about important properties of goals, while a procedural view of goals is necessary to ensure that goals can be achieved efficiently in dynamic environments. In this paper we propose a framework for goals which integrates both views. We discuss the requisite properties of goals and the link between the declarative and procedural aspects, then derive a formal semantics which has these properties. We present a high-level plan notation with goals and give its formal semantics.

Negotiation is essential in settings where autonomous agents have conflicting interests and a desire to cooperate. For this reason, mechanisms in which agents exchange potential agreements according to various rules of interaction have become very popular in recent years as evident, for example, in the auction and mechanism design community. However, a growing body of research is now emerging which points out limitations in such mechanisms and advocates the idea that agents can increase the likelihood and quality of an agreement by exchanging arguments which influence each others ’ states. This community further argues that argument exchange is sometimes essential when various assumptions about agent rationality cannot be satisfied. To this end, in this article, we identify the main research motivations and ambitions behind work in the field. We then provide a conceptual framework through which we outline the core elements and features required by agents engaged in argumentation-based negotiation, as well as the environment that hosts these agents. For each of these elements, we survey and evaluate existing proposed techniques in the literature and highlight the major challenges that need to be addressed if argument-based negotiation research is to reach its full potential.

Numerous methodologies for developing agent-based systems have been proposed in the literature. However, their application is still limited due to their lack of maturity. Evaluating methodologies' strengths and weaknesses plays an important role in improving them and in developing the "next-generation" of methodologies. This paper presents a comparison of three prominent agent-oriented methodologies: MaSE, Prometheus and Tropos. It is performed based upon an attribute-based framework which addresses four major areas: concepts, modelling language, process and pragmatics. The objectivity of the comparison is increased by including inputs from the authors of the methodologies using a questionnaire and by conducting an experimental evaluation of the methodologies.

Agent-based computing is a promising approach for developing applications in complex domains. However, despite the great deal of research in the area, a number of challenges still need to be faced (i) to make agent-based computing a widely accepted paradigm in software engineering practice, and (ii) to turn agent-oriented software abstractions into practical tools for facing the complexity of modern application areas. In this paper, after a short introduction to the key concepts of agent-based computing (as they pertain to software engineering), we characterise the emerging key issues in multiagent systems (MASs) engineering. In particular, we show that such issues can be analysed in terms of three different ‘‘scales of observation’’, i.e., in analogy with the scales of observation of physical phenomena, in terms of micro, macro, and meso scales. Based on this characterisation, we discuss, for each scale of observation, what are the peculiar engineering issues arising, the key research challenges to be solved, and the most promising research directions to be explored in the future.

Intelligent agents are a modelling paradigm, based on the notion of agents with mental states. The agent metaphor is nowadays used in many research and industry projects, and several generic agent platforms are available. Nevertheless, there is a gap between platforms concentrating on agent communication infrastructure and platforms concerned with the representation of internal agent concepts. This article presents an approach to bridge this gap: Jadex, an add-on to the widely used JADE agent platform. The add-on follows the BDI architecture, a well-known model for representing mentalistic concepts in the system design and implementation. The article provides an overview of the BDI model, and the design and realization in Jadex, as well as the integration of the add-on into the JADE agent framework.

Debugging multi-agent systems (which are concurrent, distributed, and consist of complex components) is difficult, yet crucial. We propose that the debugging process can be improved by following an agent-oriented design methodology, and then using the design artifacts in the debugging phase. We present an example of this scheme which uses interaction protocols to debug agent interaction. Interaction protocols are specified using AUML and are translated to Petri nets. The debugger uses the Petri nets to monitor conversations and to provide precise and informative error messages when protocols aren't correctly followed by the agents.