Is it possible to use "fsolve" function by using matrix inputs in order to solve the system of nonlinear equations ?

For example,

function F=myfun(x,A,B,C) {...} end

x0, A, B, C are all matrix inputs

Thank you,

[Torsten]

Call fsolve as

x=fsolve(@(x)myfun(x,A,B,C,...),x0)

This makes it possible to import additional data in "myfun".

Or what exactly is your question ?

Best wishes

Torsten.

[END]

Thank you for your answer. My problem is that when I try to solve the problem by using "fsolve" command, Matlab says "FSOLVE requires the following inputs to be of data type double". My inputs are all multidimensional matrices

[John D'Errico]

A multi-dimensional matrix usually IS of type double. However, nothing stops you from creating matrices of type uint8. fsolve cannot work with those type matrices. So you need to cast the result to doubles.

Note that often, working with integer matrices in your computations means that the computations will be quantized internally. This will make your objective function non-differentiable, so fsolve will still fail. Your objective must be differentiable, so a tiny increment in the parameters MUST result in a non-zero increment in the objective. Otherwise, fsolve will terminate immediately, since it will see a zero gradient vector.

[Torsten]

Maybe you work with symbolic matrices ? This is not allowed in fsolve...

Best wishes

Torsten.

Answer 1

[Star Strider]

What, exactly, do you want to solve?

If you have a nonlinear function of several independent variables, and you want to fit that function to data, the lsqcurvefit function might be most appropriate.

[END]

There are 2 nonlinear equations and 2 unknowns, but the parameters used in the equations are all multidimensional matrices. Consider 3D matrices as inputs

Thank you for your help

[Star Strider]

I don’t have a clear idea what you want to do, but you would probably have to write a function that used each ‘page’ of your 3D matrices as individual 2D matrices, since matrix operations (to the best of my knowledge) are only defined for 2D matrices.

Answer 2

[Rik Hati]

Hello ! Did you get the solution ? I am stucked in the same probem .