I want to use global optimization toolbox to get the gains of this control signal
Where α > 0, β > 0 ,T>0, p and q are positive odd constants, and q > p .I don't know how I can enter this constraint to optimtool especially,p and q are positive odd constants. can anyone help?
Dear Mohamed Tawfiky,
Type the following command in MATLAB command window and fill the required information;
optimtool
Further, to write necessary files see one example just by typing
help ga
in the command window.
For even more details you can search how to use optimtool in MATLAB
Dear Muhamed,
Before starting with global optimization you need to write the function as you write in the question in matlab file, also is required to put the boundary constrains to find the optimization for your case.
If you needed more help you could contact me in this email address: [email protected]
HI!
An example of the problem decision:
function outPrm = MyOptimizationFcn(x, y)
%INPUTS
%x - time
%y - measured value
%
%OUTPUTS
%outPrm - needed output parameters
global xG yG
xG = x; yG = y;%global variables for connection with "targFun"
fitTyp = fittype('here write your function. a1, a2 and so on - are optimizable parameters', 'coeff', ...
{'a1', 'a2', 'a3', 'a4'});
%example fitTyp = fittype('(a1 * x .^ 3) + (a2 * x .^ 2) + (a3 * x) + (a4 * x .^ -1)', 'coeff', ...
% {'a1', 'a2', 'a3', 'a4'});
fitOpt = fitoptions(fitTyp);
% a1 a2 a3 a4
fitOpt.Upper = [1e5, 1e5 1e10, 1e15];%put upper limits for parameters
fitOpt.Lower = [1e0, 1e0, 1e0, 1e0];%put lower limits for parameters
fitOpt.StartPoint = fitOpt.Lower + rand(1, numel(fitOpt.Upper)) .* (fitOpt.Upper - fitOpt.Lower);%random start point
cv = fit(xG, yG, fitTyp, fitOpt);%optimization by mean "optimtool"
outPrm = fminsearch(@targFun, [cv.a1, cv.a2, cv.a3, cv.a4]);%find for local minima
function tf = targFun(p)
global xG yG
%here you can enter any condition for your function
%as example:
myfun = (p(1) * xG .^ 3) + (p(2) * xG .^ 2) + (p(3) * xG) + (p(4) * xG .^ -1);
%myfun must be lower yG
nuls = 1e3 * (yG < myfun);
tf = sum(((yG - myfun) .^ 2) .* nuls);
Thank you all professors,
I built this control signal on Simulink. I want to get the best gains that make control signal minimize the error.
I wrote the objective function already. it's as following.
function J = ObjectiveF(x)
global e T B alf
T=x(1);
B=x(2);
alf=x(3);
e=sim('optry.mdl');
f= ((sum(abs(e))));
J=f;
end
then i used the optimtool ,but the results are so bad.
- Badges
- Science topic