Publisher’s version / Version de l'éditeur:
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
RuleML 2011: the 5th International Symposium on Rules: Research Based and
Industry Focused [Proceedings], 2011-07-01
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.
https://nrc-publications.canada.ca/eng/copyright
NRC Publications Archive Record / Notice des Archives des publications du CNRC :
https://nrc-publications.canada.ca/eng/view/object/?id=0c21987d-d963-444d-893d-5fb93a7e8b7b
https://publications-cnrc.canada.ca/fra/voir/objet/?id=0c21987d-d963-444d-893d-5fb93a7e8b7b
NRC Publications Archive
Archives des publications du CNRC
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Cross-community interoperation between the EMERALD and rule
responder multi-agent systems
Cross-Community Interoperation Between the
EMERALD and Rule Responder Multi-Agent Systems
Kalliopi Kravari1, Taylor Osmun2, Nick Bassiliades1 and Harold Boley2
1Dept. of Informatics, Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece
{kkravari, nbassili}@csd.auth.gr
2Inst. for Information Technology, NRC Canada, Fredericton, NB, E3B 9W4, Canada
{taylor.osmun, harold.boley}@nrc.gc.ca
Abstract. The Semantic Web will let people delegate complex behavior to in-telligent agents, which will act on behalf of their users in a variety of real-life applications. The focus here is on the semantic multi-agent systems EMERALD and Rule Responder, which can be employed to assist communities of users based on Semantic Web and multi-agent standards such as RDF, OWL, Ru-leML, and FIPA. The present work demonstrates how these multi-agent sys-tems can interoperate to automate collaboration across communities using a declarative, knowledge-based approach. A multi-step interaction scenario among agents is presented where users are considering sponsoring a symposium modeled by two communities. This scenario demonstrates the usefulness of in-teroperating between EMERALD and Rule Responder, indicating a general ap-proach to cross-community collaboration.
Keywords: rules, semantic web, intelligent multi-agent systems, gateways, EMERALD, Rule Responder, SymposiumPlanner.
1
Introduction
The Semantic Web (SW) is an evolving extension of the Web, where the semantics of information and services is well-defined, making it possible for people and machines to precisely understand Web content. Moreover, SW technologies offer interoperabili-ty and, thus, favor Intelligent Agents (IAs). Hence, the integration of multi-agent
sys-tems (MAS) with SW technology affects the use of the Web; agents are available to
traverse the Web and perform actions on behalf of their users in real-life applications. At present, a number of multi-agent systems are available; however they are typi-cally isolated, as their organizational philosophy and architecture is different and their agents usually do not share the same logic or rule representation formalism. In this work two such multi-agent systems, EMERALD [2] and Rule Responder [4], have been made interoperable through appropriately defined gateways, which could be re-used in a variety of interoperation scenarios. Each agent, whether it belongs to EME-RALD or to Rule Responder, has its own policy, a set of private rules representing its
requirements, obligations and restrictions, as well as its personal knowledge about the world.
2
EMERALD: A Multi-Agent Knowledge-Based Framework
EMERALD is a multi-agent knowledge-based framework [2], which offers flexibility, reusability and interoperability of behavior between agents, based on Semantic Web and FIPA language standards. The main advantage of this approach is that it provides a safe, generic, and reusable framework for modeling and monitoring agent commu-nication and agreements.In order to model and monitor the parties involved in a transaction, a generic, reus-able agent prototype for knowledge-customizreus-able agents (KC-Agents), consisted of an agent model (KC Model), a yellow pages service (Advanced Yellow Pages Service) and several external Java methods (Basic Java Library), is deployed. Agents that comply with this prototype are equipped with a Jess rule engine and a knowledge base (KB) that contains environment knowledge (in the form of facts), behavior patterns and strategies (in the form of Jess production rules). The use of the KC-Agents proto-type offers certain advantages, like interoperability of behavior between agents, as opposed to having behavior hard-wired into the agent’s code.
Finally, as agents do not necessarily share a common rule or logic formalism, it is vital for them to find a way to exchange their position arguments seamlessly. Thus, EMERALD proposes the use of Reasoners [3], which are actually agents that offer reasoning services to the rest of the agent community. This approach does not rely on translation between rule formalisms, but on exchanging the results of the reasoning process of the rule base over the input data. The receiving agent uses an external rea-soning service to grasp the semantics of the rulebase, i.e. the set of conclusions of the rule base. One of these Reasoners is the defeasible logic Reasoner, based on DR-DEVICE [1]. Defeasible reasoning was selected because of its simple rule-based approach for efficient reasoning with incomplete and inconsistent information.
Following the above specifications we commit to SW and FIPA standards, namely, we use the RuleML language for representing and exchanging agent policies and e-contract clauses, since it has become a de facto standard. In addition, we use the RDF model for data representation both for the private data included in agents’ internal knowledge and the reasoning results generated during the process.
3
Rule Responder
Rule Responder [4] is a tool for creating virtual organizations as multi-agent systems that support collaborative teams on the Semantic Web. It provides the infrastructure for rule-based collaboration between the distributed members of such a virtual organi-zation. Human members of an organization are assisted by semi-autonomous rule-based agents, which use Semantic Web rules to describe aspects of their owners' deri-vation and reaction logic.
Each Rule Responder instantiation employs four classes of agents, an Organiza-tional Agent (OA), Personal Agents (PAs), External Agents (EAs) and ComputaOrganiza-tional Agents (CAs). The OA represents goals and strategies shared by its virtual organiza-tion as a whole, using a global rule base that describes its policies, regulaorganiza-tions, oppor-tunities, etc. Each PA assists a single person of the organization, (semi-autonomously) acting on his/her behalf by using a local knowledge base of derivation rules defined by the person. Each EA uses a Web (HTTP) interface, accepting queries from users and passing them to the OA. Each CA can be seen as an (often low level) agent that performs an automated (computing) task.
CAs are comparable to PAs. Their output is meant to assist the OA in answering the query from the EA. They are designed to perform very specific tasks that may involve invoking services independently from the rest of the virtual organization.
The OA employs an OWL ontology as a "responsibility assignment matrix" to find a PA that can handle an incoming query. The OA uses reaction rules to send the query to this PA, receive its answer(s), do validation(s), and send answer(s) back to the EA.
4
EMERALD – Rule Responder Interoperation Gateway
EMERALD and Rule Responder (RR) were compared regarding their agent-connection topologies, their interchange principles, their used subsets of RuleML language and the role of the Prova language.Fig. 1. The EMERALD – Rule Responder gateway architecture.
The above comparison resulted in three main differences; first, the systems use different technologies (Mule and Java servlets for RR – JADE and Java agents for EMERALD), second, they use different RuleML sublanguages (Reaction RuleML for RR – DR-RuleML for EMERALD) and finally, RR has centralized management through an OA written in Prova, while in the de-centralized EMERALD architecture, a Prova reasoner is just one of the supported reasoners. Based on this analysis, (bidi-rectional) RuleML gateways between EMERALD and Rule Responder were designed and implemented. Thus, the Rule Responder EMERALD (RR – EMERALD) Gate-way was implemented as a new CA that handles an appropriate communication chan-nel. On the other hand, the EMERALD Rule Responder (EMERALD – RR) Gateway
was implemented as a new proxy agent in EMERALD, communicating directly with RR OA. Figure 1, displays the above architecture.
5
The Implemented Scenario
A scenario where an external-to-SymposiumPlanner partner (an EMERALD agent) would like to sponsor the RuleML-20XY Symposium was selected to demonstrate the EMERALD – RR Gateway (Figure 3). SymposiumPlanner (Figure 2) is a series of use cases based on the RuleML Symposium series (e.g. http://2010.ruleml.org) created with Rule Responder (RR). Using Friend of a Friend (FOAF) profiles, each chair position (general chair, panel chair, etc) has a Personal Agent (PA). Each PA has a knowledge base containing the responsibilities of the position in order to answer queries relevant to the chair's role.
Fig. 2. Rule Responder architecture for the SymposiumPlanner application.
In this scenario the partner has to decide whether or not to sponsor RuleML-20XY Symposium. The decision on the sponsoring level will be based on its personal prefe-rences, related to the benefits of each level. The latter can be obtained from the cor-responding RuleML SymposiumPlanner chair, namely the Publicity Chair.
Thus, the EMERALD agent has to communicate with the PublicityChair in the SymposiumPlanner application. First of all, it sends its query (requesting the sponsor-ing levels and their benefits) to the RRP, the Rule Responder Proxy agent (an EME-RALD agent), in order to forward it to the PA. RRP forwards the query, receives the response and returns it back to the partner. However, the decision making of the EMERALD agent is based on rules, and more specifically on defeasible logic rules.
The partner transforms the received RuleML message to RDF, in order to be used as a fact base for the rule base, which is formed in a defeasible RuleML dialect. The rule base contains its personal preferences and a link to the data that will be used (the RDF file) and it is sent to the defeasible logic reasoner (DR-Reasoner), hosted by EMERALD, in order to find out the best sponsoring level.
Fig. 3. The scenario overview.
Afterwards, the partner receives back DR-Reasoner’s response (in this case the decision was the Gold sponsoring level, among Bronze, Silver, Gold and Emerald) and sends a new query to the PublicityChair (through RRP) requesting the appropriate submission information for that level (e.g. to contact Sponsor chair by e-mail or phone). Finally, the partner is able to contact to the Sponsor chair or to continue this conversation in order to get any additional information.
Information (and related source code) about both the EMERALD – Rule Respond-er intRespond-eropRespond-eration project and the above scenario is available at the project’s site (http://lpis.csd.auth.gr/systems/EMERALDRR)
6
References
1. Bassiliades N., Antoniou G., Vlahavas I.: A Defeasible Logic Reasoner for the Semantic Web. IJSWIS, 2(1):1-41(2006)
2. Kravari K., Kontopoulos E., Bassiliades N.: EMERALD: A Multi-Agent System for Knowledge-based Reasoning Interoperability in the Semantic Web, 6th Hellenic Confer-ence on Artificial IntelligConfer-ence (SETN 2010), LNCS, Vol. 6040/2010, pp. 173-182 (2010) 3. Kravari K., Kontopoulos E., Bassiliades N.: Trusted Reasoning Services for Semantic Web
Agents, Informatica: Int. J. of Computing and Informatics, 34(4), pp. 429-440. (2010) 4. Paschke A., Boley H., Kozlenkov A., and Craig B.: Rule Responder: RuleML-Based
Agents for Distributed Collaboration on the Pragmatic Web. In 2nd ACM Pragmatic Web Conference (2007)
5. Osmun T., Smith D., Boley H., Paschke A.: Rule Responder Guide, http://ruleml.org/RuleResponder/RuleResponderGuide/
