Why do AnalysisPoints change the results of systune optimization?
I am wondering why additional analysis points change the outcome of the optimization with the systune command. The results are quite similar, but I would have expected identical results, since - as far as I understood - analysis points should not affect the system dynamics. % Define Laplace variable s = tf('s'); % Define all single transfer functions k12 = 0.5; k21 = -0.1; G11 = 2/(s^2 + 3*s + 2); G12 = k12/(s+1); G21 = k21/(s^2 + 2*s + 1); G22 = 6/(s^2 + 5*s + 6); % Concatenate MIMO system G = [G11, G12; G21, G22]; % Create tunable state-space system C = tunableSS('C', 4, 2, 2); % Define some analysi points X = AnalysisPoint('X', 2); E = AnalysisPoint('E', 2); U = AnalysisPoint('U', 2); % Closed loop WITH analysis points Gw1 = feedback(E*C*U*G, X, -1); Gw1.InputName = 'w'; Gw1.OutputName = 'y'; % Closed loop WITHOUT analysi points Gw2 = feedback(C*G, X, -1); Gw2.InputName = 'w'; Gw2.OutputName = 'y'; % Make sure the bode plots are identical before the optimiziation with % systune figure; bode(Gw1, Gw2); legend; % Specify tuning goal Rtrack = TuningGoal.StepTracking('w', 'y', 0.5/3); % Tune both closed loop systems with the same tuning goal CL1 = systune(Gw1, Rtrack); CL2 = systune(Gw2, Rtrack); % Review bode plots of the optimization results figure; bode(CL1, CL2); legend;
Kshitij Singh answered .
2025-11-20
The difference in the results of the systune optimization when including or excluding AnalysisPoint blocks is primarily due to how systune treats these blocks in the optimization process. Although AnalysisPoint blocks do not directly affect the system dynamics, they provide additional flexibility for tuning and evaluation during the optimization.
Here are the key factors that explain the difference:
When AnalysisPoint blocks are included, systune treats them as additional points where performance metrics can be evaluated. These points act as "virtual sensors" or "virtual actuators" that decouple signals for analysis purposes. This allows systune to optimize the closed-loop system in a more granular way by isolating specific signals for evaluation.
AnalysisPoint blocks, the system is optimized as a whole.AnalysisPoint blocks, systune can fine-tune the interactions between these points and the rest of the system.This additional flexibility can lead to slightly different optimal controller parameters.
AnalysisPoint blocks normalize signals for better conditioning during optimization. This normalization can affect how systune scales the internal weights and priorities of tuning goals, potentially leading to different results.
AnalysisPoint blocks: The normalization ensures better numerical stability and alignment of the optimization process.AnalysisPoint blocks: The system dynamics are treated without intermediate normalization, which might influence the optimization trajectory.When AnalysisPoint blocks are present, they allow systune to access the intermediate closed-loop transfer functions (e.g., TU→XT_{U \to X}, TE→XT_{E \to X}) explicitly. This could change the optimization outcome if these intermediate transfer functions influence the tuning goal evaluation.
Some tuning goals, like TuningGoal.StepTracking, depend on evaluating closed-loop responses. If AnalysisPoint blocks are included, they create additional signal paths and modify how the goal is evaluated. For instance:
AnalysisPoint blocks: The step tracking goal may be evaluated using decoupled paths.AnalysisPoint blocks: The goal is evaluated directly on the main signal path.These differences can alter the optimized controller parameters slightly.
To verify the exact difference caused by AnalysisPoint blocks:
getBlockValue(C)).getIOTransfer.viewGoal(CL1, Rtrack) and viewGoal(CL2, Rtrack)).AnalysisPoint blocks consistently to align the optimization process with your analysis needs.bode, step, or custom simulations.In your specific example, the inclusion of AnalysisPoint blocks adds flexibility to the optimization process, which is why the results differ slightly between the two configurations.