Research‎ > ‎Projects‎ > ‎

Reactive Mobility

The goal of the project is to simplify the development of mobile agents. The backbone of the proposal is to develop a mobility model for allowing the agent developer to delegate decisions about mobility on the run-time support of mobile agents. The model will be validated by extending the MoviLog agent programming language. In this way we will show that it is possible to automatize decisions about mobility without negativelly impacting the performance.

Introduction

A mobile agent is a software entity able to control its location on a computer network, thus it is capable of autonomously moving to the location that hosts those resources that the agent requires for performing certain computation. Mobile agents are a promissory paradigm for developing massively distributed applications, because they provide advantages on performance, scalability, fault tolerance and flexibility with respect to client/server.

Despite the clear advantages that mobile agents have shown, they have not been applied on industrial nor mission critical applications. This is mainly caused by the important development efforts involved in the adoption of the paradigm. Indeed, systems using mobile agents are more complex than stationary (non-mobile) systems, due to a simple fact: developers have to provide the code of the application plus code for managing mobility. Therefore, mobile agent systems posses an extra dimension of complexity that traditional stationary systems do not: mobility.

Another important factor that hinders the adoption of mobile agents is the usage of general purpose programming languages such as C++ or Java for its implementation. These languages do not provide specific abstractions for building agents. As a consequence, developers often have to use non-appropriate abstractions which causes more programming effort.

Multi-paradigm languages based on concepts from logic programming and object orientation have been successfully used to implement agents. These languages are a convenient choice given the requirements for agent programming. We will limit the scope of this proposal to the language JavaLog, a multi-paradigm integration between Java and Prolog, and MoviLog, an extension of JavaLog for supporting mobile agents.

The goal of the project is to simplify the development of mobile agents. The backbone of the proposal is to develop a mobility model for allowing the agent developer to delegate decisions about mobility on the run-time support of mobile agents. The model will be validated by extending the MoviLog agent programming language. In this way we will show that it is possible to automatize decisions about mobility without negativelly impacting the performance.

Publications

  • Chronos: A Multi-Agent System for Distributed Automatic Meeting Scheduling. (A. Zunino, M. Campo). Expert Systems With Applications. Elsevier Science. ISSN: 0957-4174. Ed.: J. Liebowitz. In press. 2008. (Thomson SCI). View draft.
  • Extending MoviLog for Supporting Web Services. (C. Mateos, A. Zunino, M. Campo). Computer Languages, Systems and Structures. Elsevier Science. ISSN: 0096-0551. Ed.: R. S. Ledley. Vol 31, Num. 1. pp 11-31. 2007. (Thomson SCI)
  • Reactive Mobility by Failure: When Fail Means Move. (A. Zunino, M. Campo, C. Mateos.) Information Systems Frontiers. Special Issue on Mobile Computing and Communications: Systems, Models and Applications. Springer Science+Business Media B.V. ISSN 1387-3326. Eds.: Giri Tayi and S. S. Ravi. Vol. 7, Num. 2, pp 141-154. 2005. (Thomson SCI)
  • Integrating Intelligent Mobile Agents with Web Services. (C. Mateos, A. Zunino, M. Campo.). International Journal of Web Services Research. Idea Group Inc. ISSN 1545-7362. Eds.: Liang-Jie Zhang. Vol 2, Num. 2, pp. 85-103. 2005. (Thomson SCI)
  • MoviLog: A Platform for Prolog-Based Strong Mobile Agents on the WWW (A. Zunino, M. Campo, C. Mateos). Inteligencia Artificial, Revista Iberoamericana de I.A. AEPIA. Vol. 7, num. 21, pp. 83-92. ISSN 1337-3601. Also in [2]. 2003. (Latindex)
  • Simplifying Mobile Agent Development Through Reactive Mobility by Failure (A. Zunino, M. Campo, C. Mateos). En Advances in Artificial Intelligence. Eds. Guilherme Bittencourt y Geber Ramalho. Lecture Notes in Artificial Intelligence. Springer-Verlag. Vol. 2507. ISBN 3-540-00124-7. ISSN 0302-9743. pp 163-174. 2002. (Thomson SCI)
  • Enhancing Agent Mobility Through Resource Access Policies and Mobility Policies (A. Zunino, C. Mateos, M. Campo). Proceedings of the V Encontro Nacional de Inteligencia Artificial (ENIA). XXV Congresso da Sociedade Brasileira de Computacao (SBC). pp 851-860. ISBN 85-7669-033-0. San Leopoldo, RS, Brasil. 2005.
  • Reactive Mobility by Failure (A. Zunino). Quinto Taller de Investigadores en Ciencias de la Computación (WICC 2003), pp. 646-650. Tandil, Buenos Aires. 2003.
  • MoviLog: A Platform for Prolog-Based Strong Mobile Agents on the WWW (A. Zunino, M. Campo, C. Mateos). Proceedings of the IV Argentine Symposium on Artificial Intelligence (ASAI2002), 31 Jornadas Argentinas de Informática e Investigación Operativa (JAIIO). pp. 62-73. ISSN 1666-1141. Santa Fé Septiembre 2002.