Control and Estimation of Ensembles via Structured Optimal Transport

A driving force in the development of technology, and hence society as a whole, is the automation of processes that were previously executed manually. For many tasks, this comes with an intrinsic need for sensing, estimation, and actuation-concepts that are well known in this community. Moreover, an increasing number of such applications involve large networks containing ensembles of agents that interact with other agents and the environment around them. Such physical systems, which communicate through digital communication channels and are autonomously controlled by computer-based algorithms, are called cyberphysical systems[1]. Examples that are currently being developed include self-driving cars [2], smart electric grids [3], and the Internet of Things [4]. However, there are plenty of other multiagent systems of interest, where the agents are not physical systems designed by humans. Examples of such dynamical systems include groups of animals [5], opinion dynamics among humans [6], and economic markets [7].