We consider the problem of controlling the dynamic state of each of a finite collection of targets distributed in physical space using a much smaller collection of mobile agents. Each agent can attend to no more than one target at a given time, thus agents must move between targets to control the collective state, implying that the states of each of the individual targets are only controlled intermittently. We assume that the state dynamics of each of the targets are given by a linear, time-invariant, controllable system, and develop conditions on the visiting schedules of the agents to ensure that the property of controllability is maintained in the face of the intermittent control. We then introduce constraints on the magnitude of the control input and a bounded disturbance into the target dynamics and develop a method to evaluate system performance under this scenario. Finally, we use this method to determine how the amount of time the agents spend at a given target, before switching to the next in its sequence, influences the control of the states of the entire collection of targets.

中文翻译：

---

使用可达性分析在持续监控任务中调度多个代理

我们考虑使用小得多的移动代理集合来控制分布在物理空间中的每个有限目标集合的动态状态的问题。每个智能体在给定时间最多只能关注一个目标，因此智能体必须在目标之间移动以控制集体状态，这意味着每个单独目标的状态只能间歇性地控制。我们假设每个目标的状态动态由一个线性的、时不变的、可控的系统给出，并在代理的访问时间表上开发条件，以确保在面对间歇控制时保持可控性. 然后，我们将控制输入幅度的约束和有界干扰引入目标动态，并开发了一种在这种情况下评估系统性能的方法。最后，我们使用这种方法来确定代理在切换到下一个序列之前在给定目标上花费的时间如何影响整个目标集合的状态控制。