Network systems often possess significant delays. Traditionally, Smith predictor has often been used to overcome the performance limitation imposed by the delay. In recent work, it was shown that the Multi-Scale Control scheme can also provide satisfactory performance/robustness for a system with long delay.
everything you want to know about PID tuning, see these references:
1) Aidan O'Dwyer, Handbook of PI and PID Controller Tuning Rules, 3rd edition, Imperial Collage Press, 2009.
2) SW. Sung, J. Lee, IB. Lee, Process Identification and PID Control, IEEE Press, John Wiley & Sons (Asia) Pte Ltd, 2009.
In my view, there is no "best" thing in control theory. You obtain "something" while losing "others" inevitably. The same is the word "optimal". The size, structure, communication direction (uni-, bi-) of the network and the dynamic property of each agent will determine the overall behaviors of the network system for which you want to change by a controller. None of them is clear to me from your comments.
The tuning of a PID controller is not the issue in such a problem; there are so many methods to do that. In my opinion, to set up the desired features of the network in a mathematical model is the task. For example, some people work on the congestion using the control theory.
- Badges
- Science topic
- Similar topics
- Correction