curve fitting using multiple starting points for every variable

curve fitting using multiple starting points for every variable

Given a random parametric function with n parameters used to fit
datapoints. You define you're parametric function by heigthmodel You're n
startingvalues are defined as initial values For the fitting you use
lsqcurvefit(heigthmodel,initialvalues,...
Is there a way you can give more then one startingvalue for every
parameter. So that you add another string of startingvalues and the model
will start from both startingvalues and will return the best result?
I've already setup something like this:
Heightmodel = ....
Startingvalues = a1 a2 a3 a4 a5
Startingvalues = b1 b2 b3 b4 b5
options=optimoptions(numberofvariables,2) % based on the TypicalX function.
lsqcurvefit(heightmodel,startingvalues,...,options)
It would be interesting to know if I could make this work and if so if I
could add more startingvalues. So I will have p strings of startingvalues
and numberofvariables,p
Also is there a way to present these startingvalues in a matrix form? So
that for the first variable the fitting can choose to start from both a1
or b1, for the second variable from both a2 or b2? ...
Greetings