Showing papers presented at "European Workshop on Multi-Agent Systems in 2004"

Towards active perception in situated multi-agent systems

Abstract: Modeling the environment and agent-environment relationships is not well explored in multi-agent systems, in particular not for software multi-agent systems. This paper aims to contribute with a generic model for active perception in situated multi-agent systems. Active perception enables an agent to direct its perception at the most relevant aspects in the environment, according to its current task. The model decomposes perception into three functionalities: sensing, interpreting, and filtering. The agent first senses its neighborhood through a set of selected foci, resulting in a representation. A set of perceptual laws enforces domain specific constraints on sensing. Next, the agent interprets the representation by means of descriptions, resulting in a percept. Percepts are expressions that can be understood by the internal machinery of the agent. Finally, the percept is filtered by a set of selected filters, restricting the perceived data according to specific context relevant selection criteria.

Norms in Multiagent Systems: some Implementation Guidelines

Abstract: Norms are commonly used in MAS to formally express the expected behaviour of agents in open environments. Current norm formalisms focus on the declarative nature of norms. However, in order to be implemented, norms should be translated into operational representations. In this paper we continue our work on the implementation of norms by discussing issues on norm enforcement, verifiability and defeasibility. We propose some implementation guidelines, including some mechanisms to be added in agent platforms in order to ease norm implementation.

A Multiagent System for the Reliable Execution of Automatically Composed Ad-hoc Processes

Abstract: Reference LIA-CONF-2004-008View record in Web of Science Record created on 2006-12-13, modified on 2017-05-12

A Multi Agent Based Simulator for Managing a Container Terminal

Abstract: Container terminals world wide are trying to expand capacity and increase performance at a minimum of investments. Often the container terminal operations are changing to meet increased customer demands as well as to adapt to new technologies. The increasing costs of container terminal development do justify the use of computer simulation to assist in planning and policy making. A flexible simulator which represents the marine-side operations of a container terminal with the yard operations is developed in order to test berth assignment management polices and yard stacking management policies. The simulator is designed using a Multi-Agents Systems approach. This type of simulator provides a modular and scalable method of modeling the entities in the container terminal domain. The developed simulator is called SIMPORT and is presented in the context of a case study of a major container terminal in India.

Enhancing Java Grid Computing Security with Resource Control

Abstract: This paper outlines an original Computational Grid deployment protocol which is entirely based on Java, leveraging the portability of this language for distributing customized computations throughout large-scale heterogeneous networks. It describes practical solutions to the current weaknesses of Java in the fields of security and resource control. In particular, it shows how resource control can be put to work not only as basis for load balancing, but also to increase the security and general attractiveness of the underlying economic model.

3D electronic institutions : social interfaces for e-commerce

Abstract: Artificial Intelligence Research Institute (IIIA),Barcelona, Catalonia, Spainsierra@iiia.csic.esAbstractIn this paper we propose the use of Virtual Worldsas a 3D social interface for Electronic Institutions.It is shown how the two metaphors, Virtual Worldsand Electronic Institutions are combined into a singlemetaphor, 3D Electronic Institutions, while retainingthe features and advantages of both. Its essence isto open Electronic Institutions to human users. Thestrong methodology and powerful tools provided withElectronic Institutions create the possibility of semi-automatic generation of 3D social interfaces. UsingE-Commerce as the application domain for the newmetaphor, important cues such as social interactions,establishment of long term contacts, enhanced prod-uct observation and other real world attributes thatare missing from current on-line Web-based businessenvironments can be introduced.Keywords: Virtual Worlds, Electronic Institutions,social user interfaces, multi-agent systems, E-Commerce.1 IntroductionIndividuals today are the product of a particularlymobile and entrepreneurial society. As a result, in ev-eryday business activities, including e-commerce, theindividual is socially constituted and socially situated(Solomon 2004). However, researchers point out thatsocial needs of customers play a crucial role and are ofgreat importance in E-Commerce applications Preece& Maloney-Krichmar (2003), this needs are mostlyneglected in nowadays systems. At present, most e-commerce environments and their interfaces are de-signed with the assumption that customers buyingstyle (or behaviour) is rational, without much con-sideration of the needs of customers showing an emo-tional style of buying. Some conventional e-commercesolutions already try to meet social needs of theirusers. The popular web site Friendster.com, for in-stance, assists people in finding new contacts withthe help of other people they know and trust. Theanswer to the question “Whom do you know?” ap-pears to be a key characteristic of a businessman’ssuccess.Jung & Lee (2000) propose an environment whichincorporates a spatially-organized and interactive sitemap. This site map visualizes the location of peopleand offers facilities for social interactions. Thougha good starting point, this approach does not go farenough in terms of consistency. For example, the sitemap is a separate part of the web site and not anintegral aspect of the user interface.In contrast, 3D Virtual Worlds provide a consis-tent and immersive user interface which implicitly in-corporates location awareness of other users and of-fers mechanisms for social interaction. Virtual Worldssupport to a certain extent the way humans operateand interact in the real world. Such immersive en-vironments incorporate social context, allowing cus-tomers to meet and interact with other people, andhave the potential to address the needs of both ratio-nal and emotional customers.Virtual Worlds go beyond the document and formbased interface of the World Wide Web, representingusers in a graphical form (an “avatar”, cf. (Damer1998)) and permitting them to operate, interact andcommunicate in a shared visual space, i.e. puttingthe human “in” the World Wide Web rather than“on” the World Wide Web. The underlying tech-nology supporting these worlds have been createdand hosted by variety of companies, ranging fromcomputing technology giants like Microsoft and Intel,and ending with relatively small start-ups like Ac-tive Worlds and Worlds Inc. Overall, the design anddevelopment of Virtual Worlds has emerged as a phe-nomenon shaped by the home computer user, ratherthan by the research and development in universitiesor companies. As a result, Virtual Worlds are some-what unregulated environments, which do not havethe means to enforce technological norms and ruleson their inhabitants. Hence, Virtual Worlds do notaddress the central issue of security in E-Commerce.In order to address this issue and make use of thebenefits of Virtual Worlds interface we need to in-troduce methodologies on which reliable and secureE-Commerce systems build upon. Electronic Insti-tutions, for instance, focus exactly on taking controlover security aspects. In particular, Electronic In-stitutions guarantee that participants adhere to in-stitution rules and fulfill their obligations. In gen-eral, actors in Electronic Institutions are softwareagents and practical applications of this methodol-ogy so far aimed at building societies of completelyautonomous software agents – humans did not partici-pate at all! Nevertheless the MASFIT project pointedout that humans participating in complex decisionmaking tasks quite reluctantly delegate this activityto a completely autonomous entity, (Cun´i, Esteva,Garcia, Puertas, Sierra & Solchaga 2004). This high-lights that a better understanding and modelling ofthe relationship between humans and the agents thatmake decisions on their behalf is needed.Our major objective is to bridge the gap betweenthe two metaphors, Electronic Institutions and Vir-tual Worlds, and combine them in the metaphor of 3DElectronic Institutions. Electronic Institutions enablethe specification of highly regulated and well struc-tured environments which contrast the mostly unreg-ulated nature of Virtual Worlds, cf. Figure 1.A 3D Electronic Institution is an environment thatenables human activities and participation in Elec-tronic Institutions. In order to achieve this goal weapply a three-layered framework and present a 3D so-cial user interface that allows humans to interact ef-

Efficiently Distributing Interactions between Composed Information Agents

Abstract: When information agents are organized to provide some composed functionality, their interactions can be formally represented as workflows. Traditionally, workflows are executed by centralized engines that invoke the necessary agents and collect results. If agents are clustered (e.g., based on geographic criteria), locally routing intermediary results between agents in the same cluster can be more efficient. This paper describes a mechanism that allows the execution of workflows in a distributed way. Intermediary results are routed directly from the producing service to the consuming one, minimizing the network traffic to the client. This approach is particularly useful if the client uses a mobile device with a slow or expensive network connection.

A dialogue game to agree to disagree about inconsistent information

Abstract: This paper proposes a dialogue game in which coherent conversational sequences at the speech act level are described of agents that become aware they have a disagreement and settle the dispute by agreeing to disagree when they believe insufficient propositions to resolve the situation. A dialogue game is formulated in which agents can offer information possibly resulting in non-reconcilable, mutually inconsistent belief states. These states are handled by a dialogue rule that defines when to ‘agree to disagree’.