|
|
|
|
Ontology for Autonomous Systems |
|
|
|
Written by Julia Bermejo
|
|
Wednesday, 26 October 2005 |
OASYS: Ontology for Autonomous Systems
Tesis Doctoral
Universidad Politecnica de Madrid
- Author:
- Julita Bermejo (also known as Julia Bermejo)
- Tutor:
- Ricardo Sanz
Abstract
The main goal of the thesis is to address which features should an ontology fulfil to define agents and multi-agent systems capable of being autonomous within the control systems domain, mainly regarding real-time and embedded systems.
Autonomus Agents
An intelligent agent has been defined as a computer-based system which shows properties such as autonomy, social ability and which is usually implemented using concepts related to human beings (23).
The broad concept of intelligent agent was constraint to pay special attentionto the property of autonomy. Dictionaries define the term autonomy in a general way as "condition showed by someone who does not depend on anyone under certain circumstances" (16). This generic definition has been re-defined to address science and technological domains. More precisely, autonomy has been defined asa set (System, Task, Environment) in which a System is capable of carrying outa Task in a certain Environment, mainly related to intelligent control systems(19).
From autonomy, a new term was coined: Autonomous agent. There are several definitions of such an agent in the literature. The most well-known defines autonomous agent as " a system, totally or partially situated in an environment, that perceives it, acts on it, with time, to fulfil its goal as well as to take intoaccount the effects of its actions on the future" (10).
Ontologies and Ontological Engineering
The word ontology belongs to Philosophy, meaning the study of the being and the existence (according to Aristotles definition). Its use within the Computer Science and Knowledge Engineering domains, implies the necessity to define the concepts and relationships within an specific domain. The goal is not to define what is and what it is not, but to define useful concepts for computer-based systems.
A well-known definition (within the Knowledge Engineering domain) was providedby Gruber: " an ontology is an explicit specification of a conceptualisation" (12). Lately, Ontological Engineering has been used to describe all the activities related to the development process, ontology life-cycle, methodologies, tools and languages to build Ontologies (11).
The research on Ontologies is wide. There are several methodologies (Cyc, KACTUS, METHONTOLOGY, ONIONS, PROMPT). Upper-level ontologies have been defined (CYC, SUO ); ontologies to represent knowledge (RDF, OIL, DAML+OIL, OWL, etc). Languages such as KIF, LOOM, OKBC, SHOE, OIL, OWL; tools such as Ontolingua, WebOnto, OntoEdit, Proteg-2000. To read a complete description on existing methodologies, tools and languages see (11).
Control Systems
Control systems have evolved from classical theories to new ones to address real-time and embedded features. Future control system will be characterised by being smaller, cheaper and with shorter development time, higher functionality and evolution capability. Therefore, researchers should address fields such as: agent technology, artificial intelligence, embedded systems, integration, ontologies and distributed real-time systems (17).
Computer-based control systems have traditionally been applied to process control, robotics, etc. Recent research on control topics has shown the necessity of higher autonomy. However, the higher the autonomy, the more information is required and the more adaptation capacity in uncertain environments. Therefore, conscious machines of both the system and the environment are needed (18).
Status of this work
From October 2003, preliminary work on State of the Art (or literature review)regarding:
Ontologies
- What is an ontology (existing definitions, elements and features)
- Existing ontologies
- Methodologies
- Language
- Tools for ontologies
- Ontologies for autonomous agents (addressing the concepts of autonomy, intelligence, etc)
Agents and Multi-agent systems
- What is an agent (not an unique answer, I am afraid)
- What does "autonomous" mean? (an agreement on the concept of autonomy?)
- Classification of agents
- Architectures, languages and tools
- Agents in the control systems domain
Scheduled end
July 2006.
Thesis Objectives
The researchers interest on autonomous and intelligent agents is not recent. Several European and international organisations work to standardise and to disseminate knowledge regarding agents (Agentlink (1), FIPA (8), AgentWeb(2), etc). Several architectures, languages, methodologies and tools have been developed(see (15), (23) and (24)).
Nevertheless, on the one hand autonomous agents have not always been considered to be applied the control systems domain. Recently, it has been pointed out the idoneity of using such agents (or more precisely, semi-autonomous agents) co-operating in control systems (14). Other works on agents applied to control systems can be found in (20), (21) and (22).
On the other hand, there has been several attempts to define ontologies for agents and MAS (see (9), (25), (26) and (27)).
The new approach addressed by the current thesis is the combination of three aspects: ontological engineering, autonomous agents and control systems. Up to date, such topics were addressed two by two ( ontologies for agents, agents for control systems) but in a multidisciplinary way.
The purpose of the thesis is the development of and ontology to define the concepts, relationships and architecture to be in a multi-agent system within the real-time and /or embedded control systems. The goal is to simplify the design,update and development of such agents, and as a consequence, of the control system.
The developed ontology would allow to define the requirements that agents should fulfil when applied to control systems. Other aspects to be addressed are how the ontology would allow the definition of concepts and meaning generation for the agents in the control system.
The thesis considers topics related to Cognitive Science, Consciousness, Semantics, Computer Science, etc. This multi-disciplinary and cross-domain approachwill exploit and combine the existing knwoledge on each of them.
Potential Applications
In the very long term, the definition of systems capable of being self-aware, conscious, intelligent and fully autonomous.
Documents and links
- Concepts
- A First at Concepts J. Bermejo
This document is a preliminary and not comprehensive literature on the topic of concept from definition to implementation.
- Agents
- Agents J. Bermejo.An attempt to know what an agent is, types on agents and so forth. Not finished yet, as it turned out to be an Herculean task.
- Meaning Generation andArtificial Wisdom R. Sanz, J. Bermejo, J. Escasany, R. Chinchilla and Carlos A. Garcia. Our attempt to define a new term: sapient agent.
- Conscious systems
- Machines out of the Matrix: The Emerging Science of Artificial Consciousness R. Sanz and J. Bermejo. Some general ideas on consciousness applied to control systems.
References
- Agentlink, http://www.agentlink.org
- AgentWeb, http://agents.umbc.edu
- Ashri, R. Rahwan, I. and Luck, M. (2003), Architectures for negotiating agents, Multi-agent systems and applications III, Eds. V. Marik, J. Mueller and M. Pechoucek, Lecture Notes in Artificial Intelligence, Springer, Vol. 2691,pp. 136-146.
- Botti, V. Carrascosa, C. Julian, V. and Soler, J. (1999), Modelling Agents in Hard Real-Time Environments, MAAMAW'99, pp. 63-76.
- Cranefield, S. and Purvis, M. (2001) UML-based ontology for modelling software agents, Proc. of Ontologies in Agent Systems Workshop, Agents 2001, Montreal, pp. 21-28.
- Davidsson, P. (1996), Autonomous Agents and the Concept of Concepts, PhD Thesis, Department of Computer Science, Lund University.
- De Jong, E.D. (2000), Autonomous formation of concepts and communication, PhD Thesis, Vrije Universiteit Brussel.
- FIPA, http://www.fipa.org
- Flake, S., Geiger, C. and Küster, J.M. (2001) Towards UML-based Analysis and Design of Multi-Agent Systems, International Symposium on Information Science Innovations in Engineering of Natural and Artificial Systems (ENAIS2001),Dubai, March 2001.
- Franklin, S. and Graesser, A. (1997) Is it an Agent, or just a Program?: A Taxonomy for Autonomous Agents, Proceedings of the Third International Workshop on Agent Theories, Architectures and Languages, Intelligent Agents III, Springer Verlag, pp. 21-35.
- Gómez-Pérez, A., Fernández-López, M. And Corcho, O. (2003) Ontologicalengineering: with examples from the areas of knowledge management, e-commerce and the semantic web, Springer.
- Gruber T.R. (1993) A translation approach to portable ontology specification, Knowledge Acquisition Vol. 5, No. 2, pp. 199 -220.
- Guarino N. Giaretta, P 81995) Ontologies and Knowledge bases: Towards a terminological Clarification, in Towards Very Large Knowledge Bases: Knowledge Building and Knowledge Sharing (KBKS95), University of Twente, pp. 25-32.
- Jennings, N.R. and Bussmann, S. (2003), Agent-based Control Systems, IEEE Control Systems Magazine, Vol. 23, No. 3, pp. 61-73.
- Luck, M. and d'Inverno, M. (2004), The Agent Landscape, in Understanding Agent Systems, Eds. M. d'Inverno and M. Luck, Springer, pp. 1-10.
- Real Academia Española, Diccionario de la Lengua Española
- Sanz, R. and Arzen K--E. (2003), Trends in Software and Control, IEEE Control Systems Magazine, Vol. 23, No. 3, pp 12 -15.
- Sanz, R. and Bermejo, J. (2004), Machines out of The Matrix, INDUMÁTICA2004 - INDUTEC, pp. 14-15.
- Sanz, R., Matía, F. and Galán, S. (2000), Fridges, Elephants and the meaning of autonomy and intelligence, Proceedings of the 15th IEEE InternationalSymposium on Intelligent Control (ISIC 2000), Río, Patras, Greece, 17-19 July,pp. 217-222.
- Van Breemen A.J.N. (2001), Agent-based Multi-controller systems, PhD Thesis, University of Twente.
- Velasco, J.R., González, J.C., Magdalena, L. and Iglesias C.A. (1996) Multiagent-based control systems: a hybrid approach to distributed process control, Control Engineering Practice, Vol. 4, No. 6, pp. 839-845.
- Vools, H. (2000) Intelligent Agents for Supervision and Control: A Perspective, , Proceedings of the 15th IEEE International Symposium on IntelligentControl (ISIC 2000), Río, Patras, Greece, 17-19 July, pp. 339-344.
- Wooldridge, M. and Jennings, N.R. (1995), Intelligent Agents: Theory and Practice, Knowledge Engineering Review, Vol. 10, No. 2, pp. 15-152.
- Zarnekow, R. (1998), Fundamental Concepts of Intelligent Software Agents in Intelligent Software Agents: Foundations and Applications, Springer, pp. 19-34.
- (2001) Proceedings of the Workshop on Ontologies in Agent Systems (OAS01), 5th International Conference on Autonomous Agents Montreal, Canada, May 29, Eds. S. Cranefield, T. Finin and S. Wilmott.
- (2003) Proceedings of the IJCAI-03 Workshop on Ontologies and Distributed Systems, Acapulco, August 9, 2003, Eds. F. Giunchiglia, A. Gómez-Pérez, A. Pease, Y. Sure and S. Wilmott.
- (2003) Proceedings of the Workshop on Ontologies in Agent Systems (OAS03), 2nd International Joint Conference on Autonomous Agents and Multi-Agent Systems, Melbourne, Australia July 15 2003, Eds. S. Cranefield, T. Finin, V. Tamma and S. Wilmott.
Copyright ž 2004 ASLab.
Universidad Politecnica
de Madrid
|
Last update:
Nov. 2004
All rights
reserved |
.
| | |
