The Fabricare system: a multi-agent-based scheduling prototype

Abstract

Several trends in society in general, and manufacturing in particular, have changed the way business was made in the last decades of the 20th century, setting new requirements for companies and individuals. The research question being addressed in this paper is concerned with the ability to build computer-supported manufacturing systems able to cope with current and future requirements. For this matter, a hypothesis based on the holonic and multi-agent paradigms is proposed. The paper describes a holonic architecture for manufacturing enterprises and a prototype system (named Fabricare) for manufacturing orders scheduling based on that architecture. A negotiation mechanism called ‘contract net with constraint propagation protocol’ was developed for regulating the interaction between holons in the system. This protocol also implements a negotiation-driven scheduling procedure.

Keywords:

• Holonic manufacturing systems
• multi-agent systems
• scheduling

Acknowledgments

The authors would like to thank FCT–‘Fundação para a Ciência e Tecnologia’ (Portuguese Foundation for Science and Technology) and programmes POCTI and FEDER for their support in R&D projects. Acknowledgements are also given to the Scientific Council of ISEP/IPP and to FLAD – Fundação Luso-Americana para o Desenvolvimento’ (Portuguese-American Foundation for Development). Additionally, we acknowledge the European Commission which supported some ESPRIT and IST programmes in which we were involved.

Paulo Sousa studied computer science at Instituto Politécnico do Porto (Politechnic Institute of Porto Portugal – ISEP-IPP from 1994 to 1995 with a specialization in Industrial Informatics. He worked for three years as an application developer for a Portuguese software house in the field of electronic archive, database retrievals and component building. In 1996 became an assistant professor at ISEP-IPP. In 1998, he successfully concluded a post-graduation on ^‘Distribution Systems, Computer Architectures and Computer Communications’ at Universidade do Minho (University of Minho, Portugal), and in 2002 achieved his PhD degree in Computer Science at Universidade do Minho with his work on Holonic Manufacturing Systems. He is currently an active member of the ISEP's Research Group on Knowledge Engineering and Decision Support (GECAD, http://www.gecad.isep.ipp.ptl/) and the technical contact point for the ISEP-IPP node of AgentLink – European Network of Excellence in Agent-Ifased computing. His main research interests are Computer Graphics, Distributed Intelligent Systems and Enterprise Application Architectures. More information can be found at http://www.dei.isep.ipp.pt/∼psousa.

Carlos Ram was born in São Paulo, Brazil, in 1963. He got his BSc (licenciatura) degree in Electrical Engineering in 1986, and a PhD degree in Electrical and Computer Engineering in 1993, both from the University of Porto, Portugal. Currently he is Co-ordinator Professor at the Institute of Engineering–Polytechnic of Porto (ISEP-IPP). He is Director of the ISEP Research Group on Knowledge Engineering and Decision Support (GECAD, http://www.gecad.isep.ipp.pt/). He was responsible for 14 R&D projects, and supervised 16 MSc thesis and seven PhD works, and is supervising more on going PhD work. He has published more than 274 scientific papers. His main areas of interest are Artificial Intelligence and Manufacturing Systems, namely involving Decision Support, Knowledge-based Systems, and Multi-agent Technologies.

José Neves was born in Santa Cristina de Malta–Vila do Conde, Portugal, in 1948. He got his BSc (licenciatura degree in Chemical Engineering in 1976, at the University of Coimbra, Portugal, MSc and PhD degrees in Computer Science in 1981 and 1983, both from the University of Heriot-Watt, in Edinburgh, Scotland. Currently he is Full Professor at the University of Minho (UM), in Portugal. He is responsible for the UM Research Group on Artificial Intelligence and co-ordinates several R&D projects involving Artificial Intelligence basic research and application. He applied with success several Artificial Intelligence technologies like Logic Programming, Neural Networks, Constraints, Evolutionary - Programming, Multi Agent Systems, Negotiation and Argumentation in areas like Medicine, Law and Legal Domain, Electronic Commerce, Manufacturing, etc. He has had more than 150 scientific papers pubslished.