Adaptive Management Systems for Control Tasks in Cyber-physical Systems

提出了在网络物理系统中最优适应当前环境的控制任务管理方法。利用弹性任务模型，我们将实时计算系统中能量的降低和控制性能的提高之间的权衡描述为一个优化问题。然后，我们得到了计算其最优解的算法。接下来，我们提出了使用多版本任务模型和跳过作业来避免过载情况的两种方法。最后，我们设计了一个控制任务，用于在线搜索多个智能体协同工作的最优位置。

We proposed management methods of control tasks which are adaptive to the current environment optimally in cyber-physical systems. Using an elastic task model, we formulated a trade-off between the energy reduction in real-time computing systems and the improvement of control performances as an optimization problem. Then, we obtained an algorithm for computing its optimal solution. Next, we proposed two methods to avoid overload situations using a multi-version task model and job skipping. Finally, we designed a control task to search an optimal position on-line in a multi-agent system where each agent works cooperatively.