[Scilab-users] Least squares problem call to fmincon - HELP

[David Brant]

Many thanks for getting back Stéphane. I tried your suggestion with the
gradient option, but got a problem:

 !--error 26 
Too complex recursion! (recursion tables are full)
at line       2 of function fun called by :  
at line       2 of function fGrad called by :  
sample_fGrad=fGrad(x0)
at line     649 of exec file called by :    
init=fun(x0)

The minimalist call also returned a problem:

y=numderivative(fun,x)
                       !--error 4 
Undefined variable: 10000000000000G*1000100
at line      29 of function flag1 called by :  
at line      16 of function fGrad1 called by :  
init=fun(x0)
at line     842 of exec file called by :    
API Error:
in error: bad call to getVarAddressFromPosition! (1rst argument).

Probably a stupid error on my part, but cannot spot it!

Code:
function z=fun(p)
	z=DAT(:,2)-p(1)*exp(p(2)*DAT(:,1))-p(3)*ones(DAT(:,1))
endfunction

//function dz=dfun(p)
//	var=exp(p(2)*DAT(:,1))
//	dz=[-var..
//	-p(1)*DAT(:,1).*var..
//	-ones(var)]
//endfunction

function f=costf(p)
	f = norm(fun(p))^2;
endfunction

//function g=grad(p)
//	g = 2*dfun(p)'*fun(p);
//endfunction 

DAT=[0 0;0 1;1 1;2 1.5;2 2];
p0=[0 0 0];

xopt = fmincon(costf,p0,[],[])

//options = list("GradObj",grad)
//xopt = fmincon(costf,p0,[],[],[],[],[],[],[],options)

fmincon.sce <http://mailinglists.scilab.org/file/t490392/fmincon.sce>  



--
Sent from: http://mailinglists.scilab.org/Scilab-users-Mailing-Lists-Archives-f2602246.html