In order to realize the potential for reliability and maintainability offered by distributed computer control systems in real-time process control applications, it is essential to have flexible methods for incorporati...
详细信息
In order to realize the potential for reliability and maintainability offered by distributed computer control systems in real-time process control applications, it is essential to have flexible methods for incorporating redundancy. A practical architecture based on local-area network technology which supports this incorporation of redundancy is proposed. A set of classification criteria for interprocess communication primitives is given and the applicability of the various proposals which have appeared in the literature is analysed. An approach based on the validity time of messages is outlined. This concept permits a unified treatment of message transfer protocols and transparent redundancy in the application software.
The objective of this paper is to examine the problem stated as follows: In what way the production process parameters and the distributedcomputer system elements influence the DCCS performance. The analysis of the p...
详细信息
The objective of this paper is to examine the problem stated as follows: In what way the production process parameters and the distributedcomputer system elements influence the DCCS performance. The analysis of the production process and the discussion of the distributedcomputer system have been presented herein. The simulation model of the DCC3 has been described and discussed. The simulation results have been given. Some design aspects of the DCCS have been presented.
The complexity of decentralised control, large scale systems and distributed processing are combined in the design of distributed computer control systems. A general view of the problems that arise in the design of di...
详细信息
The complexity of decentralised control, large scale systems and distributed processing are combined in the design of distributed computer control systems. A general view of the problems that arise in the design of distributed computer control systems and some of the present approaches to solve them are presented. Particular emphasis in the cross fertilization of computer science and control theory is made.
When a real time control algorithm is to be executed in several computational facilities (multiple processor, multi-processor, network etc.), the problem of how to decompose the control algorithm arises. The decomposi...
详细信息
When a real time control algorithm is to be executed in several computational facilities (multiple processor, multi-processor, network etc.), the problem of how to decompose the control algorithm arises. The decomposition should consider the concurrent process synchronization, interprocess communication and performance problems. In this paper a single formal tool that allows one to treat different aspects of the above problems is developed: Time Valued Petri Nets. Time Valued Petri Nets are a simple tool for modelling, decomposing and analysing centralized real time control algorithms for parallel processing. A great variety of aspects can be analysed with this tool: decomposition, data flow, concurrency, error propagation, execution time, interprocess communication, trade off of system management requirements and number of processes, assignment of processes to processors and number of processors required. The application of Time Valued Petri Nets to a life size multivariable model reference adaptive control algorithm has shown the ease and power of its use.
暂无评论