Iterative Learning Control

Abstract

We explore the problem of identification and control of living cell populations. We describe how de novo control systems can be interfaced with living cells and used to control their behavior. Using computer controlled light pulses in combination with a genetically encoded light-responsive module and a flow cytometer, we demonstrate how in silico feedback control can be configured to achieve precise and robust set point regulation of gene expression. We also outline how external control inputs can be used in experimental design to improve our understanding of the underlying biochemical processes. AbstractModel predictive control (MPC) has become the standard for implementing constrained, multivariable control of industrial continuous processes. These are processes which are designed to operate around nominal steady-state values, which include many of the important processes found in the refining and chemical industries. The following provides an overview of MPC, including its history, major technical developments, and how MPC is applied today in practice. Possible future developments are provided. AbstractInformation and communication complexity of a networked control system identifies the minimum amount of information exchange needed between the decision makers (such as encoders, controllers, and actuators) to achieve a certain objective, which may be in terms of reaching a target state or achieving a given cost threshold. This formulation does not impose any constraints on the computational requirements to perform the communication or control. Both stochastic and deterministic formulations are considered. AbstractThe concept of "information structures" in decentralized control is a formalization of the notion of "who knows what and when do they know it." Even seemingly simple problems with simply stated information structures can be extremely hard to solve. Perhaps the simplest of such unsolved problem is the celebrated Witsenhausen counterexample, formulated by Hans Witsenhausen in 1968. This entry discusses how the information structure of the Witsenhausen counterexample makes it hard and how an information-theoretic approach, which explores the knowledge gradient due to a nonclassical information pattern, has helped obtain insights into the problem. AbstractMulti-agent systems are encountered in nature (animal groups), in various domains of technology (multi-robot networks, mixed robot-human teams) and in various human activities (such as dance and team athletics). Information exchange among agents ranges from being incidentally important to crucial in such systems. Several systems in which information exchange among the agents is either a primary goal or a primary enabler are discussed briefly. Specific topics include power management in wireless communication networks, data-rate constraints, the complexity of distributed control, robotics networks and formation control, action-mediated communication, and multi-objective distributed systems. AbstractThe notion of input to state stability (ISS) qualitatively...