Application of holonic part-oriented control architecture to a machining line

Article type: Research Article

Authors: Tharumarajah, A.

Affiliations: CSIRO Manufacturing Science and Technology, Locked Bag 9, Preston, Vic. 3072, Australia. Tel.: +61 3 9662 7752; Fax: +61 3 9662 7851; E-mail: [email protected]

Abstract: Holonic manufacturing professes a system of highly distributed, autonomous and cooperating entities called holons. The modularity of such systems would naturally lend itself to greater flexibility in both construction and operation and would equip them to better cope with day-to-day disturbances and otherwise costly re-configuration. Architecture to design holonic manufacturing in a shop floor should consider many aspects of system design and operation. Broadly, these include definition of appropriate manufacturing holon classes, flexibility in specification and reconfiguration of holons and their associations, communication services, coordination of activities and so on. Overshadowing these aspects is that a holonic system should be able to produce a solution to the overall production problem as efficiently as the current systems. This requirement is often neglected to the detriment of sub-optimal production performance. In fact, holonic or similar heterarchical structures, while being robust in the face of unexpected disturbances may suffer from myopic decision-making, lack of conflict resolution and stability that are not conducive to sustaining good delivery and utilization performance of the shop floor. The holonic part-oriented architecture proposed here attempts to address these concerns. It defines holons and their associations taking predominantly a part lifecycle view of manufacturing and hence a control on value added tasks. Additionally, it considers the required functional distinction and the need for spatial associations. Functional distinction among holons of a class not only simplifies control, but also allows for better cohesion of decisions. Social associations support optimization of utilization performance by exploiting the conjunctive relationships that exists among resources. It is believed that these two, combined with a focus on part lifecycle, would help to achieve a superior performance of a holonic manufacturing system. This architecture is applied to develop specifications of a holonic model of an actual machining line and the performance of the model is demonstrated through simulation.

DOI: 10.3233/ICA-2002-9303

Journal: Integrated Computer-Aided Engineering, vol. 9, no. 3, pp. 219-233, 2002