What is the meaning of real values of unknown coefficients at Multi Start using lsqcurvefit?

Im working on solving lsqcurvefit problems and I got a suspicion that the solutions were not optimized for the fitting results.

Thus, I have tired to control the optimoptionis and found out Multi Start using lsqcurvefit could be effective.

I checked the example of Multi Start on this url https://jp.mathworks.com/help/gads/multistart-using-lsqcurvefit-or-lsqnonlin.html

However, I counldnt understand the meaning of preal which indicated real values of unknown coefficients at this example.

Here my questions :

1. What is the meaning of preal at Multi Start example?

2. Should I set this 'preal' invariably when I fit my original function to my experimental data sets?

Answer 1

[Matt J]

In that example, preal are simulated true parameter values used to generate fake data. It is not part of the work flow when working with real experimental data. You will not normally know the true parameters.

[HYOJAE LEE]

Thank you for your quick answer.

Now I got one more question.

What is the meaning of multistart iterations?

Could you explain the meaning of multistart iterations or Local solver call ?

And how do I set this value? Can I set this value randomly? like 500 or 10000?

[Matt J]

That question might be good to post in a different thread. If your original posted question has been answered, though, please click "Accept".