Synthesis of Optimal Controllers for Cyber Physical Systems

Cyber Physical Systems (CPSs) are usually composed of a set of controllers which monitor and control a physical environment through a set of sensors, actuators and communication devices. Failure of these systems in satisfying their operational goals can cause damage to the system being controlled or to people depending on them. Therefore it is important to ensure that these systems meet their requirements. On the other hand, complexity of CPSs makes the design process hard, tedious and prone to errors. To deal with these problems, a plausible approach is to systematically build the system where the correctness follows from construction. In this framework, we consider controller synthesis problem: given a model of the system and a specification, the goal is to construct a controller such that synthesized system satisfies its specification. We focus upon a subset of these problems where the goal is to find an optimal controller, i.e. a controller that optimizes a quantitative measure.We review some of the recent works on synthesis of optimal controllers. We show how these works, though similar in context, differ significantly in the model they assume for system, the form of specification considered, the techniques used to solve the problem and optimization criterion.