The ArguGRID Project
| Francesca Toni Imperial College UK |
| Editor: Francesca Toni & Fariba Sadri |
Download: PDF
1 Introduction
The ArguGRID project1 aims at developing a grid-based platform populated by rational decision-making agents that are associated with service requestors/providers and users. Within agents, logic-based argumentation [2,3,4,11] is used to support decision making, taking into account (and despite) the often conflicting information that these agents have, as well as the preferences of users, service requestors and providers. Argumentation is also intended to support the negotiation between agents [8,10], on behalf of service requestors/providers/users. This negotiation takes place within dynamically formed virtual organisations. The agreed combination of services, amongst the agents, can be seen as a complex service within a service-centric architecture [1]. We intend to validate this overall approach by way of industrial application scenarios.We have chosen to focus on e-business applications as we believe that they will benefit from a grid-based realisation, while at the same time illustrating and making use of the “semantic” techniques envisaged by ARGUGRID. These applications are intended to provide context for the project and to guide the development of formal models, their implementation, and subsequent experiments. Concretely, the chosen e-business scenarios are (see [5] for details):
1. e-Procurement applications and e-Marketplaces,
2. e-Business for Earth Observation applications, and
3. the problem of business planning and outsourcing to new countries
2 ArguGRID: an Overview
ArguGRID aims to:- develop argumentation-based foundations for the GRID, populated by rational decision-making agents within virtual organisations.
- incorporate argumentation models into service-centric architecture.
- develop underlying platform using peer-to-peer computing and overlay networks.
- validate the ArguGRID approach by way of industrial application scenarios.
Fig. 1. ArguGRID perspective
The envisaged system to support this vision is schematically described in Fig. 2.
Fig. 2. The ArguGRID envisaged system
Each service requestor/provider and each user is associated with one or more agents. Agents use argumentation for negotiating on behalf of service requestors/providers/users. Users can provide input to agents, in terms of their objectives (what they expect to achieve from the service composition performed by the agents) and preferences (either for the specific objectives, or, more generally, as a generic profile of the user).
Agents negotiate with one another by using argumentation to support their decision making and communication processes. Negotiation takes place within dynamically created and maintained virtual organizations, envisaged as societies of agents whereby interaction is regulated by social norms and/or protocols. The outcome of negotiation results into a contract, understood, at the agent level, as a task allocation (in terms of provision of resources/services) to agents. In particular, this contract may include a workflow description [6], that needs to be appropriately executed, for example by a workflow execution engine. In the case that a workflow results from the negotiation, we adopt a concrete service-centric architecture, instance of this general vision, whereby InforSense KDE system will allow to provide input from the user and will be responsible for executing the workflow [1]. The agents and the service centric architecture rely upon an underlying infrastructure implemented using a peer-to-peer approach [9].
Agents in ArguGRID
Agents in ArguGRID are defined using logic-based argumentation [4,11] and realized within the MARGO system2 [4] that is built on top of the CaSAPI system3 [3]. CaSAPI is a Prolog implementation of assumption-based argumentation [12,13,14,2] under several semantics, and in particular the semantics of preferred extensions, grounded extensions and ideal extension for argumentation [2], respectively generalizing the semantics of partial stable models, well-founded model and ideal models for logic programming. The specification and implementation of agents allows them to identify suitable goals and plans to achieve them, making use of preferences and despite incomplete and conflictual information.Conclusions
The project has concluded its first year and is currently facing the following challenges- full representation of abstract and concrete workflows within argumentative agents
- definition of a model to support Virtual Organisations, based upon the use of protocol representation in LCC [15] and argumentation-based negotiation [8]
- modeling of contract negotiation by argumentative agents
- use a model of trust to filter interactions
- full implementation of the ArguGRID platform, incorporating communicating agents, peer-to-peer technology and workflow authoring and execution tools.
References
1. V. Curcin, M. Ghanem, Y. Guo, K. Stathis and F. Toni. Building next generation Service-Oriented Architectures using argumentation agents. Proc. 3rd International Conference on Grid Services Engineering and Management (GSEM 2006), September 20062. P.M. Dung, P. Mancarella, F. Toni. Computing ideal sceptical argumentation. Artificial Intelligence, Special Issue on Argumentation in Artificial Intelligence (to appear in September/October 2007)
3. D. Gaertner, F. Toni. CaSAPI: a system for credulous and sceptical argumentation. Proc. LPNMR-Workshop on Argumentation and Non-Monotonic Reasoning (ArgNMR07), May 2007
4. Maxime Morge and Paolo Mancarella. The hedgehog and the fox. An argumentation-based decision support system. Proc. Fourth International Workshop on Argumentation in Multi-Agent Systems (ArgMAS 2007), May 2007
5. T. Stournaras and D. Dimitrelos and A. Tabasco and J.A. Barba and D. Pedrazzani and M.J. Yagüe and T.C. An and P.M. Dung and N.D. Hung and V. D. Khoi and P. M. Thang. e-Business application scenarios. ArguGRID deliverable D.1.2, September 2007
6. J. McGinnis, S. Bromuri, V. Urovi, K. Stathis, Automated Workflows Using Dialectical Argumentation, German e-Science Conference, Germany, 2007.
7. K. Stathis, G.K. Lekeas and C. Kloukinas, Competence checking for the global e-service society using games, Engineering Societies in the Agents World (ESAW06), Springer, 2007.
8. S. Modgil and J. McGinnis. Towards Characterising Argumentation Based Dialogue in the Argument Interchange Format, ArgMAS 2007, Hawai US, 2007.
9. K. Stathis, S. Kafetzoglou, S. Pappavasiliou, S. Bromuri., Sensor Network Grids: Agent Environment combined with QoS in Wireless Sensor Networks. Proceedings of the 3rd International Conference on Autonomic and Autonomous Systems (ICAS'07), Athens, GR, 2007.
10. T. Miller, P. McBurney, J. McGinnis, K. Stathis. First-Class Protocols for Agent-Based Coordination of Scientific Instruments. Proceedings of the 5th International Workshop on Agent-based Computing for Enterprise Collaboration: Agent-Oriented Workflows and Services (ACEC 07), Paris, France, 2007.
11. F. Toni, Assumption-based argumentation for selection and composition of services.Proceedings of the 8th International Workshop on Computational Logic in Multi-Agent Systems (CLIMA VIII)
12. A. Bondarenko, P.M. Dung, R.A. Kowalski, F. Toni, An abstract, argumentation-theoretic approach to default reasoning, Artificial Intelligence 93(1-2) pp 63-101
13. Y. Dimopoulos, B. Nebel, F. Toni, On the Computational Complexity of Assumption-based Argumentation for Default Reasoning, Artificial Intelligence 141, pp 57-78, October 2002
14. P.M. Dung, R.A. Kowalski, F. Toni, Dialectic proof procedures for assumption-based, admissible argumentation, Artificial Intelligence 170(2), February 2006, pp 114-159
15. D. Robertson. Multi-agent Coordination as Distributed Logic Programming. In Proceedings of the International Conference on Logic Programming, Sant-Malo, 2004
1 http://www.argugrid.eu/
2 http://margo.sourceforge.net/.
3 http://www.doc.ic.ac.uk/~dg00/casapi.html