Safe, Plug and Play, Multi Agent Dynamic Systems

从无人驾驶汽车，到纳米卫星网络和复杂的生物网络，现代世界有许多多智能体动态系统的例子，需要仔细协调和控制才能正确执行。在许多情况下，这些系统是使用基于直觉、计算机模拟和经验测试的设计建立起来的。然而，显然需要推进对此类系统的基本理解：（i）需要推导出可验证的整体动态系统属性，以确保在以前未设想的情况下的性能;（ii）不仅要了解固定配置的稳定系统行为，而且还要了解灵活配置（如分裂、合并和变形）的稳定系统行为

From driverless cars, to networks of nano satellites, and complex biological networks, the modern world has many examples of multi agent dynamic systems that need careful coordination and control to perform correctly. In many cases, these systems are built up using designs based on intuition, computer simulations and empirical testing. However, there is a clear need to advance the fundamental understandings of such systems: (i) Verifiable overall dynamic system properties need to be derived to give assurance of performance in situations not previously envisaged; (ii) It is also critical to understand stable system behaviours not just with fixed configurations, but with agile configurations such as splitting, merging, and morphing