A model to predict the performance of the contract net approach to distributed scheduling in a discrete parts manufacturing environment is presented. This extends previous work by explicitly considering the computational resource constraints both in the scheduling system and its underlying communication network. The model is used to demonstrate that the main constraint on the operation of such a scheduling system is the time taken by the machines to process the task descriptions which are broadcast to the machines by task managers. This constraint places restrictions on the type of task descriptions which can be used successfirlly in such systems.
1: IntroductionAlthough considerable advances have been made in manufacturing technology, the integration of individual technologies into a wider computer integrated manufacturing (CIM) system has been complicated by the lack of any generally accepted model of how a CIM system should operate. The definition of, and the interfaces between CIM control modules, and the design of the CIM mformation and communication systems are all currently areas of research [l]. In addition, there is a general need to be able to predict the performance of any implementation of a CIM design, to ensure that the overall system will work as intended.The work described here focuses on the predicted performance of an approach to job scheduling w i h n a &strib-uted CIM system, whch uses negotiation between machme controllers and a task manager to determine job allocation. The interaction between this distributed control application and the local area network used to support the application is modelled by means of a detailed simulation. The computational resources which are present in the system are considered in a way which allows the likely constraints on the overall system operation to be evaluated.
2: Control architecturesCurrently, the interaction between CIM control modules often follows a hierarchml model, with the number of control levels used, and the functions assigned to each level depending on the product and the organisational structure of a company [2][3][4]. Such herarchcal arrangements restrict a control module to communicate only with modules in adjacent control levels, and with only one higher level module.An altemative control pardgm whch advocates the use of less restricted communication paths between modules is the heterarclucal archtecture [5,6]. An example of such a structure is the use of negotiation as a method of resource allocation in complex manufacturing situations, where it has shown promising results [7-111.The negotiation techmque followed here is the contract net approach [12], in whch a task manager with jobs to be performed communicates a description of a job to the system workstations or machmes over a local area network. The machnes whlch are able to perform the task then reply with bids for the work, and the task manager awards the job to the best bidder when the bid evaluation deadline is reached. In the system studied here, bids are based on the estimated ...
The electronics clusters in Singapore and Taiwan have grown largely as a result of foreign direct investment. It is shown that foreign electronics companies in Singapore and Taiwan manufacture with processes close to the state-of-the-art and have considerable process and product design responsibilities. It is established that the products designed are not the core technology for the companies concerned. When compared to the development of the Scottish sites of foreign electronics companies, the sites in Singapore and Taiwan compare favourably in terms of time to obtain design responsibilities and the size of R & D (research and development) teams. National initiatives in Singapore and Taiwan that support the move to R & D at foreign sites are described.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.