Extremum seeking (ES) is a non-model based optimization algorithm that employs a gradient estimate adding zero-mean perturbations.
What do mean "non-model" and "zero-mean perturbations"?
What does mean "non-model"?
0
$\begingroup$
optimization
stochastic-processes
1 Answers
1
non-model seems to mean that you estimate derivatives to use for gradient-based optimisation from dynamic output feedback rather than from analysis of a model
zero-mean perturbations seems to mean that you introduce perturbations to the input in order to get some varying output, with these perturbations having zero mean so you stay close to an optimal extreme once you have found it