Feedback Control for Computer Systems

Chapter 12. Exploring Control Systems Through Simulation

The next several chapters describe—in the form of case studies—a number of control problems and also show how they can be solved using feedback mechanisms. The case studies are treated via simulations; the simulation code is available for download from the book’s website.

The ability to run simulations of control systems is extremely important for several reasons.

• The behavior of control systems, specifically of feedback loops, can be unfamiliar and unintuitive. Simulations are a great way to develop intuition that is required to solve the kind of practical problems that arise in real-world control systems.

• Extensive experimentation on real production systems is often infeasible for reasons of availability and cost. Even when possible, experiments on real systems tend to be time-consuming (the time scale of many processes is measured in hours or even days); if they involve physical equipment, they can be outright dangerous.

• Implementing controllers, filters, and other components in a familiar programming language can bring abstract concepts such as “transfer functions” to life. In this way, simulations can help to make some of the theoretical concepts more concrete.

• The parts of the simulation outside the controlled system—that is, the components of the actual control loop proper—will carry over rather directly from the simulation to a “real” implementation.

Finally, it is unlikely that any control system will be deployed into production ...