Computing and Communication Quality of Service for Distributed Time-Scaled Simulation in Heterarchical Manufacturing Control

Abstract

The authors present experimental results and empirical models that characterize the computing and communication workload for distributed time-scaled simulation used for distributed arrival time control (DATC) algorithms. In DATC application, real-time discrete event scheduling is achieved in a highly distributed manner by part entities (work pieces) that continuously control their arrival times at the queues for various resources based on feedback obtained from fast time-scaled distributed simulation. This study focuses on a range of systems of up to 180 parts and 16 machines. It focuses on real-time scheduling in heterarchical manufacturing systems using time-scaled distributed simulation. The quality (nearness to optimality) of schedules of such a system is limited by the speed and accuracy of the distributed simulations. In this paper designed experiments are used to characterize the Quality of Service (QoS). Based on experiments with time-scaled simulation for DATC algorithms, a regression model is built and is used to predict the latency and jitter on the network. The experimental results and regression model will enable systematic assessment of the QoS requirements and design of computing and communication architectures for the next generation of distributed manufacturing control systems.