A model for communication in the distributed evaluation of a control strategy

We consider the problem of the evaluation of a given control strategy, when this strategy is a function of measurements that are arbitrarily distributed among various locations. We propose a state-transition model for the parametrization of the communication involved in the computation of the resulting control action. Then we use dynamic programming over the state space to find an optimal information transfer strategy between the sites, according to a given cost model comprising communication and processing costs.