0
$\begingroup$

I have used model predictive control strategy in order to optimize a linear discrete SISo model.

For example:

$x(t+1)=-2x(t)+4u(t)$

Where I want to regulate my state $x(t)$ at the desired point (a simple regulating problem).

In every step, I predict $20$ future samples such that MPC optimizes next $20$ time instances (prediction horizon$=20$) and obtains $20$ responses for control variable(control horizon $=20$) then based on RHC technique, I just apply $10$ samples to model. I know both control horizon and prediction horizon are $20$ samples. but what about applied samples ($10$ samples) what can I name my applied samples?

$\endgroup$
  • $\begingroup$ It's not called anything particular, and I would not try to invent some special name to this. It's just MPC but with the small change that you don't update already after one sample as often as one normally does. $\endgroup$ – Johan Löfberg Apr 25 '19 at 20:01
  • $\begingroup$ Just a remark: Using acronyms without writing them out at the first usage is a very bad communication style. $\endgroup$ – MachineLearner Apr 25 '19 at 20:45
0
$\begingroup$

based on an article (Fig.2):

https://doi.org/10.3390/en11030631

control time steps is a word to define the time between control updates and iterative receding horizon optimization.

$\endgroup$
0
$\begingroup$

I found another term clearly to name control sequences in this case: manipulated control sequence means some consecutive sequences of decision variables which are ready to be applied to the system.

$\endgroup$
  • $\begingroup$ control time steps = control sampling time $T_s$, manipulated variable = control variable $\boldsymbol{u}(t)$, more formulation at here. $\endgroup$ – Arash Jul 7 '19 at 8:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.