Stochastic sequential sampling for identification and control of distributed systems

Stochastic decentralized control of distributed systems is studied from theoretical and algorithmic point of view. Decentralization is formalized by imposing conditional independence assumptions in the centralized control problem. However, local models and aims are in general incompatible with this structure and a suitable projections must be found. A range of probabilistic approaches has been proposed, leading to analytically intractable results. The aim of this project is to explore suitability of stochastic sampling techniques to this task.