How to use "ga" (genetic algorithm) for curve fitting

Hi every body.

I have two available vectors and . I want to fit the function on Γ which , , , , , and should be identifed to fit the function. So, I have Γ and x already existing in my workspace. I used the genetic algorithm as bellow to find , , , , , and .

 Fun=@(p1,w1,x,phi1,p2,w2,phi2)(p1*sin(w1*x+phi1)+p2*cos(w2*x+phi2));
 Cost=@(p1,w1,phi1,p2,w2,phi2)norm(Gamma-Fun(p1,w1,x,phi1,p2,w2,phi2));
 coef=ga(Cost,6)

but things go wrong when I run the code:

Not enough input arguments.

Error in @(p1,w1,phi1,p2,w2,phi2)norm(Gamma-Fun(p1,w1,x,phi1,p2,w2,phi2))

Error in createAnonymousFcn>@(x)fcn(x,FcnArgs{:}) (line 11)
fcn_handle = @(x) fcn(x,FcnArgs{:});

Error in makeState (line 52)
            firstMemberScore = FitnessFcn(state.Population(initScoreProvided+1,:));

Error in gaunc (line 45)
state = makeState(GenomeLength,FitnessFcn,Iterate,output.problemtype,options);

Error in ga (line 399)
            [x,fval,exitFlag,output,population,scores] = gaunc(FitnessFcn,nvars, ...

Caused by:
    Failure in initial user-supplied fitness function evaluation. GA cannot continue.

Can anybody help me how I may solve this problem with genetic algorithm?

Answer 1

[Walter Roberson]

cost_wrapper = @(x) Cost(x(1),x(2),x(3),x(4),x(5),x(6));
coef = ga(cost_wrapper, 6)

[Majid Shabani]

Thank you so much for your reply. Can you please let me know what is the different between "Cost" and "Cost_warpper" ?

All the best

[Walter Roberson]

cost_wrapper accepts a vector of 6 values and extracts each of the 6 values and passes them one-by-one to Cost.

All of the Mathworks optimizers require functions that accept vectors of values, not individual parameters.

[Majid Shabani]

Great explanation. Thank you again.