Optimization: find model parameters comparing with experimental data

Hi, I have a model (*.m) written in Matlab with several parameters. I would like to find the optimal parameters that will give me the minimum error compared to the experimental data.

Basically, I have to minimize this function (let's call it "error"):

error=|mymodel(a,b,c,d....)-expData|

"mymodel" and "expData" are just 1-D arrays. Actually they are 2-D arrays (let's call the columns "x"and "y") but I have done an interpolation of the most dense array ("mymodel" in this case), so that the I get the value of "y" for the 2 arrays at the same value of "x".

How can I do that? I have seen many tutorials but I don't quite understand. I have to make a function and use the optimization toolbox on that function, but how I can define the parameters to change?

Thank you

Answer 1

[Torsten]

Reading the description of lsqcurvefit should help:

http://de.mathworks.com/help/optim/ug/lsqcurvefit.html

Best wishes

Torsten.