nach oben

2009 | Buch

Kapitel lesen Erstes Kapitel lesen

Engineering Societies in the Agents World X

10th International Workshop, ESAW 2009, Utrecht, The Netherlands, November 18-20, 2009. Proceedings

herausgegeben von: Huib Aldewereld, Virginia Dignum, Gauthier Picard

Verlag: Springer Berlin Heidelberg

Buchreihe : Lecture Notes in Computer Science

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

The 10th international workshop “Engineering Societies in the Agents’ World” (ESAW 2009), was held in Utrecht, The Netherlands, during November 18–20, 2009. In the tradition of its predecessors, ESAW 2009 was committed to the idea of multi-agent systems (MAS) as highly interconnected societies of agents, paying particular attention to the social aspects, methodologies and software infrastructures that tackle the emergent complexities of MAS. The idea for the ESAW workshop series was born 10 years ago, in 1999, among the members of the working group on “Communication, Coordination and Collaboration” of AgentLink, the 1st European Network of Excellence on Agent-Based Computing, out of a critical discussion about the general mi- set of the agents community. Central to this discussion is the need for proper consideration of systematic aspects of MAS, acknowledging the importance of a multi-disciplinary approach, that takes into account the social, environmental and technological perspectives. These issues that are as actual today as they were in 1999, which is con?rmed by the steady interest in the ESAW workshop series that previous editions took place in: – Berlin, Germany, 2000 (LNAI 1972) – Prague, Czech Republic, 2001 (LNAI 2203) – Madrid, Spain, 2002 (LNAI 2577) – London, UK, 2003 (LNAI 3071) – Toulouse, France, 2004 (LNAI 3451) – Kusadasi, Turkey, 2005 (LNAI 3963) – Dublin, Ireland, 2006 (LNAI 4457) – Athens, Greece, 2007 (LNAI 4995) – Saint-Etienne, France, 2008 (LNAI 5485) This10thworkshopwasdevotedtothediscussionoftechnologies,methodologies and models for the engineering of complex applications based on MAS, and broughttogetherresearchersandcontributionsfrombothwithinandoutsidethe agents’?eld–fromsoftwareengineering,distributedsystems,socialsciences,and

Inhaltsverzeichnis

Frontmatter

Part 1: Invited Talks

The Immergence of Norms in Agent Worlds

Abstract

In this paper, after a short review of the dichotomous view of norms usually seen as either regular behaviors or obligations issued by authorities, norms are proposed to be defined as recognized, represented and reasoned upon prescriptive commands. A normative agent architecture – EMIL-A – is presented and shown to account for a complex bidirectional dynamics of norms as social phenomena that emerge because and to the extent that they immerge in the agents’ minds. Simulations run using EMIL-A will be discussed to illustrate the advantages of the present treatment of norms, over either side of the dichotomy.

Rosaria Conte, Giulia Andrighetto, Marco Campennì

Thinking Integral: How to Build Complex Systems That Live with People and Exhibit Collective Intelligence

Abstract

Multi-agent systems have been proposed for the development of complex software systems as a way to handle complexity (Jennings 2001). It has also been advocated that basic agent oriented systems do not have the power to cope with large software and that it is crucial to use organizations centered multi-agent systems (OCMAS) (Ferber et al. 2004 ; Dignum 2009).

Jacques Ferber

Part 2: Self-organisation

A Space-Based Generic Pattern for Self-Initiative Load Balancing Agents

Abstract

Load-Balancing is a significant problem in heterogeneous distributed systems. There exist many load balancing algorithms, however, most approaches are very problem specific oriented and a comparison is therefore complex. This paper proposes a generic architectural pattern for a load balancing framework that allows for the plugging of different load balancing algorithms, reaching from unintelligent to intelligent ones, to ease the selection of the best algorithm for a certain problem scenario. As in complex network environments there is no “one-fits-all solution”, also the integration of several different algorithms shall be supported. The presented pattern assumes autonomous agents and decentralized control. It can be composed towards arbitrary network topologies, foresees exchangeable policies for load-balancing, and uses a black-board based communication mechanism to achieve high software architecture agility. The pattern has been implemented and first instantiations of it with three algorithms have been benchmarked.

Eva Kühn, Vesna Sesum-Cavic

A Goal-Oriented Approach for Modelling Self-organising MAS

Abstract

Autonomous software agents provide a promising solution to the needs of decentralised networked systems, able to adapt their behaviour in a complex and dynamically changing environment.

Current agent-oriented software engineering methodologies tend to focus on different levels to realise such a self-adapting behaviour, namely the agent individual level and the global system level. The first requires to design a goal-directed agent behaviour, the second to design agents able to optimize their coordination with other peer agents in the organization, giving rise to system-level adaptation.

In this paper we propose to extend a goal-oriented engineering methodology to deal with the modelling of organisations that are able to self-organise in order to reach their goals in a changing environment. To deliver on this aim, we combine Tropos4AS, an extension of TROPOS for adaptive systems, with concepts, guidelines and modelling steps from the ADELFE methodology, which provides a bottom-up approach for engineering collaborative multi-agent societies with an emergent behaviour.

The resulting MAS has self-adaptation properties, having agents that are able to change their behaviour according to changes in the environment, and having organisations that adapt themselves to changing needs. The approach is illustrated by modelling a collaborative multi-agent system for conference management.

Mirko Morandini, Frédéric Migeon, Marie-Pierre Gleizes, Christine Maurel, Loris Penserini, Anna Perini

Engineering Agent Organisations in a Business Environment

Abstract

Motivated by demands from the commercial world for software systems that can assist in the reorganisation of processes for the purpose of reducing business complexity, we discuss the benefits and challenges of the multi-agent approach. We concentrate on the engineering aspects of large scale multi-agent systems and begin our exploration by focusing on a real world example from the call centre industry. The critical call routing process seems appropriate and useful in presenting our ideas and provides a good starting point for the development of agent organisations capable of self-management and coordination. The main contributions of this work can be summarised as the demonstration of the value of agent organisational models that do not replicate the typical hierarchical structures observed in human organisations and that a quite basic peer-to-peer structure produces very similar performance indicators to a mature simulator that uses conventional techniques, suggesting further improvements may readily be realized.

Dimitris Traskas, Julian Padget

Peer-to-Peer Overlay Network Based on Swarm Intelligence

Abstract

As the number of information in the Internet constantly increases and the complexity of systems rapidly grows, locating and manipulating complex data has become a difficult task. We propose a self-organizing approach that combines purely decentralized unstructured peer-to-peer (P2P) with space based computing in order to effectively locate and retrieve information from a network. The approach is inspired by swarm intelligence, is distributive and autonomous. As the scalability is a common open issue both for unstructured P2P networks and for coordination models, our approach successfully copes with that by using a biologically inspired multi-agent system. Benchmarks demonstrate powerful query capabilities with a good scalability.

Vesna Sesum-Cavic, Eva Kühn

Part 3: Software-Engineering and Architectures

Agent Architectures for Compliance

Abstract

A Normative Multi-Agent System consists of autonomous agents who must comply with social norms. Different kinds of norms make different assumptions about the cognitive architecture of the agents. For example, a principle-based norm assumes that agents can reflect upon the consequences of their actions; a rule-based formulation only assumes that agents can avoid violations. In this paper we present several cognitive agent architectures for self-monitoring and compliance. We show how different assumptions about the cognitive architecture lead to different information needs when assessing compliance. The approach is validated with a case study of horizontal monitoring, an approach to corporate tax auditing recently introduced by the Dutch Customs and Tax Authority.

Brigitte Burgemeestre, Joris Hulstijn, Yao-Hua Tan

Incorporating BDI Agents into Human-Agent Decision Making Research

Abstract

Artificial agents, people, institutes and societies all have the ability to make decisions. Decision making as a research area therefore involves a broad spectrum of sciences, ranging from Artificial Intelligence to economics to psychology. The Colored Trails (CT) framework is designed to aid researchers in all fields in examining decision making processes. It is developed both to study interaction between multiple actors (humans or software agents) in a dynamic environment, and to study and model the decision making of these actors. However, agents in the current implementation of CT lack the explanatory power to help understand the reasoning processes involved in decision making. The BDI paradigm that has been proposed in the agent research area to describe rational agents, enables the specification of agents that reason in abstract concepts such as beliefs, goals, plans and events. In this paper, we present CTAPL: an extension to CT that allows BDI software agents that are written in the practical agent programming language 2APL to reason about and interact with a CT environment.

Bart Kamphorst, Arlette van Wissen, Virginia Dignum

Programming Organization-Aware Agents

A Research Agenda

Abstract

Organizational notions such as roles, norms (e.g., obligations and permissions), and services are increasingly viewed as natural concepts to manage the complexity of software development. In particular in the context of multi-agent systems, agents are expected to be organization-aware, i.e., to understand and reason about the structure, work processes, and norms of the agent organization in which they operate. In this paper, we analyze which kinds of reasoning an agent should be able to do to function in an organization. We categorize these kinds of reasoning with respect to several dimensions, and distinguish three general approaches on how these might be integrated in existing agent programming languages. Through this, we provide a research agenda on what needs to be addressed when developing techniques for programming organization-aware agents.

M. Birna van Riemsdijk, Koen Hindriks, Catholijn Jonker

Energy Trade-Offs in Resource-Constrained Multi-Agent Systems

Abstract

Sensor networks and mobile ad hoc networks are two types of open system with resource constraints, in which functionality may be compromised by a lack of resources. In this work, we investigate adaptive algorithms for power-challenged computing networks. Using simulations, we have studied how self-organization can be used to trade off energy for (acceptable) accuracy to improve longevity in a sensor network; and how perceived threat and information sensitivity can be used to trade-off energy for (acceptable) security risk in an ad hoc network.

Hugo Carr, Jeremy Pitt, Anthony Kleerekoper, David Blancke

Part 4: Social Aspects of Agent Societies

Engineering Social Reality with Inheritance Relations

Abstract

In systems based on organisational specifications a reoccurring problem remains to be solved in the disparity between the level of abstractness of the organisational concepts and the concepts used in the implementation. Organisational specifications (deliberately) abstract from general practice, which creates a need to relate the abstract concepts used in the specification to concrete ones used in the practice. The prevailing solution for this problem is the use of counts-as statements. However, current implementations of counts-as view the relations expressed in this notion as static ontological classifications, which presents problems in dynamic environments where the meaning of abstract concepts can change over time. This limitation has already been solved in complex formal theoretical investigations, but the results of that study are far too complex to make a practical implementation. This paper investigates the limitations of current implementations of counts-as, and proposes a more flexible implementation based on the use of inheritance relations.

Huib Aldewereld, Sergio Alvarez-Napagao, Frank Dignum, Javier Vázquez-Salceda

Determining the Trustworthiness of New Electronic Contracts

Abstract

Expressing contractual agreements electronically potentially allows agents to automatically perform functions surrounding contract use: establishment, fulfilment, renegotiation etc. For such automation to be used for real business concerns, there needs to be a high level of trust in the agent-based system. While there has been much research on simulating trust between agents, there are areas where such trust is harder to establish. In particular, contract proposals may come from parties that an agent has had no prior interaction with and, in competitive business-to-business environments, little reputation information may be available. In human practice, trust in a proposed contract is determined in part from the content of the proposal itself, and the similarity of the content to that of prior contracts, executed to varying degrees of success. In this paper, we argue that such analysis is also appropriate in automated systems, and to provide it we need systems to record salient details of prior contract use and algorithms for assessing proposals on their content. We use provenance technology to provide the former and detail algorithms for measuring contract success and similarity for the latter, applying them to an aerospace case study.

Paul Groth, Simon Miles, Sanjay Modgil, Nir Oren, Michael Luck, Yolanda Gil

Trust Based Evaluation of Wikipedia’s Contributors

Abstract

Wikipedia is an encyclopedia on which anybody can change its content. Some users, self-proclaimed “patrollers”, regularly check recent changes in order to delete or correct those which are ruining articles integrity. The huge quantity of updates leads some articles to remain polluted a certain time before being corrected. In this work, we show how a multiagent trust model can help patrollers in their task of controlling the Wikipedia. To direct the patrollers verification towards suspicious contributors, our work relies on a formalisation of Castelfranchi & Falcone’s social trust theory to assist them by representing their trust model in a cognitive way.

Yann Krupa, Laurent Vercouter, Jomi Fred Hübner, Andreas Herzig

Evolutionary Role Model for Multi-Agent Systems

Abstract

In sociology, the role concept is deeply researched to predict activities of human organizations and theorized with many sub-theories. In the same direction, multi-agent system researchers use the role concept to model and program the agents behaviours, cooperations. But there is an important point missed out by the MAS researchers: evolution of the organization. In this paper, by inspiring from the efforts in sociology, we propose an evolutionary role model for coping with the evolution of the role-based multi-agent systems.

Erdem Eser Ekinci, Oğuz Dikenelli

Part 5: Organisation and Autonomy

Replication Based on Role Concept for Multi-Agent Systems

Abstract

Replication is widely used to improve fault tolerance in distributed and multi-agent systems. In this paper, we present a different point of view on replication in multi-agent systems. The approach we propose is based on role concept. We define a specific “fault tolerant role” which encapsulates all behaviors related to replication-based fault tolerance in this work. Our strategy is mainly focused on replicating instances of critical roles in the agent organization. However, while doing this, we simply transfer the critical role and the fault tolerant role to appropriate agents. Here, the fault tolerant role is responsible for coordination between replicated role instances (replicas). Moreover, our approach is flexible in terms of fault tolerance since it is possible to easily modify existing behaviors of the “fault tolerant” role, remove some of its behaviors, or include new behaviors to it due to its characteristic architecture.

Sebnem Bora, Oguz Dikenelli

Knowledge Management in Role Based Agents

Abstract

In multi-agent system literature, the role concept is getting increasingly researched to provide an abstraction to scope beliefs, norms, goals of agents and to shape relationships of the agents in the organization. In this research, we propose a knowledgebase architecture to increase applicability of roles in MAS domain by drawing inspiration from the self concept in the role theory of sociology. The proposed knowledgebase architecture has granulated structure that is dynamically organized according to the agent’s identification in a social environment. Thanks to this dynamic structure, agents are enabled to work on consistent knowledge in spite of inevitable conflicts between roles and the agent. The knowledgebase architecture is also implemented and incorporated into the SEAGENT multi-agent system development framework.

Hüseyin Kır, Erdem Eser Ekinci, Oguz Dikenelli

Balancing Organizational Regulation and Agent Autonomy: An MDE-Based Approach

Abstract

The deployment of agent societies —as complex systems— in dynamic and unpredictable settings brings forth critical issues concerning their design. Organizational models have been advocated to specify open systems in dynamic environments in order to accomplish the need to represent regulating structures explicitly and independently from acting components (or agents). Despite the fact that several frameworks have been proposed for the specification of organizational models, it is still a matter of design choice how to balance between regulative design and component flexibility.

We propose a design framework, discussing the advantages of having different degrees of abstraction at organizational level in the development of agent societies. That is, we illustrate how the design properties impact the flexibility of run-time systems to cope with context changes. We adopt the OperA software engineering methodology to deal with the organizational model specification, and the Model Driven Engineering (MDE) mechanisms to map concepts between different design models.

Loris Penserini, Virginia Dignum, Athanasios Staikopoulos, Huib Aldewereld, Frank Dignum

Cooperative Sign Language Tutoring: A Multiagent Approach

Abstract

Sign languages can be learned effectively only with frequent feedback from an expert in the field. The expert needs to watch a performed sign, and decide whether the sign has been performed well based on his/her previous knowledge about the sign. The expert’s role can be imitated by an automatic system, which uses a training set as its knowledge base to train a classifier that can decide whether the performed sign is correct. However, when the system does not have enough previous knowledge about a given sign, the decision will not be accurate. Accordingly, we propose a multiagent architecture in which agents cooperate with each other to decide on the correct classification of performed signs. We apply different cooperation strategies and test their performances in varying environments. Further, through analysis of the multiagent system, we can discover inherent properties of sign languages, such as the existence of dialects.

İlker Yıldırım, Oya Aran, Pınar Yolum, Lale Akarun

Assistance Layer in a P2P Scenario

Abstract

Usually, MAS design and implementation involves a coordination model that structures agent interactions and an infrastructure in charge of enacting it. We propose the term Coordination Support to denote the services offered by this infrastructure. Such services can be grouped in different layers. We propose an additional Assistance layer devoted to assist coordination rather than just to enable it. This layer is illustrated by means of a Peer-to-Peer sharing network (P2P) scenario, so that the benefits of our proposal can be empirically evaluated.

Jordi Campos, Maite López-Sánchez, Marc Esteva

Navigational Web-Interfaces from Formal Tropos Specification

Abstract

This paper presents a method of building executable and interactive application interface prototypes from requirements. The specification of the requirements uses i* and Formal Tropos languages.

Komminist Weldemariam

Part 6: Demonstrations

ALIVE: A Framework for Flexible and Adaptive Service Coordination

Abstract

There is a large body of research on software services, but the issues of communication and dynamic reconfiguration have received little attention, as have adaptation to environment and dynamic combination of service building blocks into new applications. Here, we present the approach of the FP7 alive project to the use of formal models of coordination and organisation mechanisms to deliver a flexible, high-level means to describe the structure of interactions between services in the environment. Our aim is to create a framework for services engineering for “live” open systems of active services. We propose to build on the current activities in service-oriented engineering by defining three levels: (i) An organisational level models the organisational structure of executing and interlinked services and the context around them. (ii) A coordination level provides flexible ways to model interaction between the services. (iii) These two levels connect with existing (semantic) Web services, which contain semantic descriptions to make components aware of their social context and of the rules of engagement with other services.

J. S. C. Lam, W. W. Vasconcelos, F. Guerin, D. Corsar, A. Chorley, T. J. Norman, J. Vázquez-Salceda, S. Panagiotidi, R. Confalonieri, I. Gomez, S. Hidalgo, S. A. Napagao, J. C. Nieves, M. Palau Roig, L. Ceccaroni, H. Aldewereld, V. Dignum, F. Dignum, L. Penserini, J. Padget, M. De Vos, D. Andreou, O. Cliffe, A. Staikopoulos, R. Popescu, S. Clarke, P. Sergeant, C. Reed, T. Quillinan, K. Nieuwenhuis

An Organisational Adaptation Simulator for P2P Networks

Abstract

Organisational centred multi-agent systems (MAS) have proved to be effective to regulate agents’ activities. Nevertheless, population and/or environmental changes may lead to a pour fulfilment of the system’s purposes, and therefore, adapting the whole organisation becomes key. This paper presents a MAS simulator devoted to test organisations with self-adaptation capabilities in P2P scenarios. More specifically, our simulator implements different sharing P2P methods –some of them with self-adaptation– and so, it can be used as a testbed for comparing them.

Jordi Campos, Marc Esteva, Maite López-Sánchez, Javier Morales

PreSage-MS: Metric Spaces in PreSage

Abstract

We consider adaptation in open systems, i.e. systems without global objects or common objectives. There are three related issues: how to make the degrees of freedom (DoFs) transparent to all agents, how to define a ‘fair’ process for performing adaptation, and how to retain some control over the adaptation to avoid, for example, undesirable configurations. We represent the specification DoFs in terms of a metric space, and define, in a uniform and consistent way, a mechanism for ‘moving’ between points in the metric space which is both ‘fair’ to the agents and avoids unacceptable moves or points in the space. This approach is demonstrated by the platform PreSage-MS, which allows a designer to specify and animate an adaptive open multi-agent system in terms of a metric space and norm-governed rules for ‘moving’ in that space.

Hugo Carr, Alexander Artikis, Jeremy Pitt

Normative Multi-Agent Organizations

A Programming Language and Its Interpreter

Abstract

Multi-agent systems are viewed as consisting of individual agents whose behaviors are regulated by organization artifacts. This abstract presents a programming language, which is designed to implement norm-based organization artifacts, by means of an example and explains the execution behavior of the language interpreter.

Mehdi Dastani

Hybrid Teams in Virtual Environments: Samurai Joins the Training Team

Abstract

This paper demonstrates a virtual environment where mixed human-agent teams are used for team-skills training.

Jurriaan van Diggelen, Tijmen Muller, Karel van den Bosch

Joint Activity Testbed: Blocks World for Teams (BW4T)

Abstract

This demonstration will be the presentation of a new testbed for joint activity. The domain for this demonstration will be similar to the classic AI planning problem of Blocks World (BW) extended into what we are calling Blocks World for Teams (BW4T). By teams, we mean at least two, but usually more members. Additionally, we do not restrict the membership to artificial agents, but include and in fact expect human team members. Study of joint activity of heterogeneous teams is the main function of the BW4T testbed.

Matthew Johnson, Catholijn Jonker, Birna van Riemsdijk, Paul J. Feltovich, Jeffrey M. Bradshaw

Backmatter

Titel: Engineering Societies in the Agents World X
herausgegeben von: Huib Aldewereld
Virginia Dignum
Gauthier Picard
Verlag: Springer Berlin Heidelberg
Electronic ISBN: 978-3-642-10203-5
Print ISBN: 978-3-642-10202-8
DOI: https://doi.org/10.1007/978-3-642-10203-5