Performance evaluation of real-time decision-making architectures for computer-integrated manufacturing systems

Abstract A general method is proposed for the performance evaluation of a decision-making architecture for computer-integrated manufacturing systems. A decision-making architecture broadens the concept of a control architecture by integrating control, communication and database functions. A modular modeling methodology is developed that captures these features and is applicable to an arbitrary computer-integrated manufacturing architecture. The model is based on generalized stochastic Petri nets and leads to a quantitative evaluation of such performance measures as response time, average utilization of a particular system component, average queue length, etc. The net result is a design tool that can be used to make tradeoffs among the system parameters. The proposed technique is demonstrated using several real-time decision-making architectures. Several general conclusions are drawn from this investigation. Finally, a Petri net model reduction method is presented for this problem and used to compare the original performance evaluation results with those obtained from the simplified models.