Task assignment and scheduling for open real-time control systems

Abstract

A problem for task assignment and scheduling on a network of processors is formulated and solved for open real-time control systems. In order to ensure smooth operation and good performance of open real-time control systems, one must analyze the problem of task assignment and scheduling during the conceptual system design stage. For this type of applications, we propose use of a performance index called the control latency, a weighted sum of feedback, command, and monitoring latencies. Given a set of tasks for a specific control application, the execution time of each task, and intra/inter-processor communication latencies, we have developed an optimal task assignment and scheduling algorithm by minimizing this performance index. Since this problem is NP-hard, we have employed a branch-and-bound algorithm to efficiently search for an optimal task assignment while maintaining task schedulability. A prototypical example of open-architecture control for CNC machines is presented to illustrate the good performance of the proposed algorithm.