Use function handle in which past values are used

Hello,

I have the input and output data of a real system:

xdata = U_set; % setted voltage
ydata = U_meas; % measured voltage

My goal is to use the function lsqnonlin to apply the Levenberg-Marquardt method and get the estimated parameter vector for the following assumed input-output-relationship:

x()... parameters

x... input from xdata

y... model output

There is an example in which the function fun doesn't contain past values of the model or past values of the input https://de.mathworks.com/help/optim/ug/lsqnonlin.html

xdata = [0.9 1.5 13.8 19.8 24.1 28.2 35.2 60.3 74.6 81.3];
ydata = [455.2 428.6 124.1 67.3 43.2 28.1 13.1 -0.4 -1.3 -1.5];
fun = @(x)x(1)*exp(x(2)*xdata)-ydata; 
x0 = [100,-1];
options = optimoptions(@lsqnonlin,'Algorithm','trust-region-reflective');
x = lsqnonlin(fun,x0,[],[],options)

How can I implement a function, in which past values of xdata or ydata can be used to use my assumed model in fun?

Best regards

KB

[Vladimir Sovkov]

You can program it as a separate function using loops, and convert it to optimization expression with the help of the call function fcn2optimexpr from Matlab Optimization Toolbox.

You can also try our package Optimizer at https://sourceforge.net/projects/optimizer-sovkov/

[KB]

I am using the free Matlab version for students and haven't access to the Optimization Toolbox.

But I solved this with an external function, which is called by the function handle.

Thank you four your help :)