RECONFIGURABLE PRODUCTION CONTROL SYSTEMS: BEYOND ADACOR Paulo Leitão 1 , João Mendes 2 , Armando W. Colombo 3 , Francisco Restivo 2 1 Polytechnic Institute of Bragança, Quinta Sta Apolónia, Apartado 1134, 5301-857 Bragança, Portugal, pleitao@ipb.pt 2 Faculty of Engineering of University of Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal, {joao.mendes,fjr}@fe.up.pt 3 Schneider Electric - HUB & Globalization of Technology, Steinheimer Str. 117, 63500 Seligenstadt, Germany, armando.colombo@ de.schneider-electric.com Abstract: In the recent evolution of production control systems, the emergence of decentralized systems capable of dealing with the rapid changes in the production environment better than the traditional centralized architectures has been one of the most significant developments. The agent-based and holonic paradigms symbolize this approach, and ADACOR holonic control architecture is a successful example of such a system. In this paper, authors discusses the current challenges and the way to go in the direction of new, reconfigurable, evolvable and ubiquitous systems, able to respond to current production environment demands and variability. Copyright © 2007 IFAC. 1. INTRODUCTION Production control systems have evolved dramatically during the last few years. One of the most significant facts is the emergence of decentralized systems capable of dealing with the rapid changes in the production environment better than the traditional centralized architectures. The quest for agility and re-configurability requires a new class of production control systems, characterized by: - A community of distributed and intelligent building blocks, designated by control units. - Each control unit is autonomous, having its own objectives, knowledge and skills, and encapsulating intelligent functions; however, none of them has a global view of the system. - Global decisions (e.g. scheduling and diagnosis) are obtained by more than one control unit, i.e. control units need to work together to reach a production decision. - Control units should exhibit some emergent behavior, such as to adapt to changes without external intervention. - Control units representing mechatronic devices, such as sensors or robots, are part of the production control system architecture. The agent-based and holonic paradigms symbolize this new approach, and ADACOR (ADAptive holonic COntrol aRchitecture for distributed manufacturing systems) [1], as others (see e.g. PROSA [2], HCBA [3] and Bussmann [4]), is a successful example of such a system. ADACOR deals with the re-configurability in manufacturing systems by introducing an adaptive production control system that evolves dynamically between a more hierarchical and a more heterarchical control architecture, based in self-organization and learning capabilities embedded in individual holons. Re-configurability, that is the ability of the system to dynamically change its configuration, usually to respond to dynamic changes in its environment, e.g. a new production scenario, assumes a key role in the new generation of production control systems, providing the way to achieve a rapid and adaptive response to change, which is a key enabler of competitiveness. Starting with an overview of how re-configurability is supported by ADACOR architecture, this paper presents the current challenges and the way to go in the direction of new, reconfigurable and ubiquitous systems, able to integrate networked production resources to respond to the variability of production scenarios beyond those that were envisaged at design time. For this purpose, the paper introduces the guidelines for the new generation of re-configurable production systems and discusses how to implement these systems, pointing out the benefits of combining multi-agent systems with service-oriented architectures.