Solution hit upper bound in lsqnonlin

Hi,

I am fitting my data using the lsqnonlin function. The equation that I need to fit to the data is quite complex with 5 unknowns. I set lower and upper boundaries given that some parameters to be biologically corrent must be within a range of values. What I've noticed is that my solution hits the upper boundary. My data are images thus each pixel has a values. Many pixels have the value of the upper boundary. What it means when the solution hits the upper limit? Can I do anything to improve the fitting?

Answer 1

[Matt J]

It means the solver thinks it can improve the fit if you remove the upper bound. If that shouldn't happen it could be due to bad data, a bad model function, or a bug in the code for the model function.