
192
■
Bio-Inspired Computing and Networking
Steam temperature
Set point
Penalty
Square insensitivity
zone
Linear insensitivity zone
510
0
10
20
30
520 550 560 570
Square + linear penalty
Only linear penalty
No penalty
540530
Figure 7.8 Example of penalty function for a steam temperature.
Global model
optimization
Stochastic
optimization
Start
SILO knowledge is sufficient
SILO knowledge is not sufficient
Steady process state and
SILO knowledge is not sufficient
Unsteady process state or
SILO knowledge is sufficient
Mixed model
optimization
Figure 7.9 Simplified scheme of layered optimization algorithm.
in this layer is a starting point for optimization in layers that use a process ...