how to solve this error? Objective function is returning undefined values at initial point. FSOLVE cannot continue.

x0=[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0];
opts=[];
x=fsolve(@fun,x0,opts)

y = 15×1

-40 -100 NaN NaN NaN NaN NaN NaN NaN NaN

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Error using fsolve (line 207)
Objective function is returning undefined values at initial point. FSOLVE cannot continue.

function y=fun(x)
H=7.9;
D=0.3145; 
P=100;
V10=0;
h=0.79;
e1=0.4;
Q=0;
R=32131;%单位kN·m²
p0=0.368*187.5*1.67*((x(3)*D^2/0.0078).^0.333);
p1=0.368*4392.2/16*1.67*((x(4)*D^2/0.0078).^0.333);
p2=0.368*5784.4/16*1.67*((x(5)*D^2/0.0072).^0.333);
p3=0.368*7176.6/16*1.67*((x(6)*D^2/0.0086).^0.333);
p4=0.368*8568.8/16*1.67*((x(7)*D^2/0.0047).^0.333);
p5=0.368*562.5*2*((x(8)*D^2/0.0064).^0.333);
p6=0.368*562.5*6.6*((x(9)*D^2/0.0017).^0.333);
p7=0.368*562.5*20.38*((x(10)*D^2/4.694898747058535e-04).^0.333);
p8=0.368*562.5*2.22*((x(11)*D^2/0.0057).^0.333);
p9=0.368*562.5*1.8*((x(12)*D^2/0.0072).^0.333);
p10=0.368*562.5*6.08*((x(13)*D^2/0.0018).^0.333);
%a=R(i-1)
a=R;
%b=-2R(i-1)-2R(i-1)+Qh^2
b=-4*R+Q*h^3;
%c=R(i-1)+4R(i)+R(i+1)-2Qh^2+Epyih^4
c0=6*R-2*Q*h^2+p0/x(3)*h^4;
c1=6*R-2*Q*h^2+p1/x(4)*h^4;
c2=6*R-2*Q*h^2+p2/x(5)*h^4;
c3=6*R-2*Q*h^2+p3/x(6)*h^4;
c4=6*R-2*Q*h^2+p4/x(7)*h^4;
c5=6*R-2*Q*h^2+p5/x(8)*h^4;
c6=6*R-2*Q*h^2+p6/x(9)*h^4;
c7=6*R-2*Q*h^2+p7/x(10)*h^4;
c8=6*R-2*Q*h^2+p7/x(11)*h^4;
c9=6*R-2*Q*h^2+p9/x(12)*h^4;
c10=6*R-2*Q*h^2+p10/x(13)*h^4;
%-2R(i)-2R(i+1)+Qh^2
d=-4*R+Q*h^3;
%e=R(i+1)
e=R;
y=[R*(x(2)-2*x(3)+x(4))-P*e1;...
    R/2*(x(1)-2*x(2)+2*x(4)-x(5))-P;...
    a*x(1)+b*x(2)+c0*x(3)+d*x(4)+e*x(5);...
    a*x(2)+b*x(3)+c1*x(4)+d*x(5)+e*x(6);...
    a*x(3)+b*x(4)+c2*x(5)+d*x(6)+e*x(7);...
    a*x(4)+b*x(5)+c3*x(6)+d*x(7)+e*x(8);...
    a*x(5)+b*x(6)+c4*x(7)+d*x(8)+e*x(9);...
    a*x(6)+b*x(7)+c5*x(8)+d*x(9)+e*x(10);...
    a*x(7)+b*x(8)+c6*x(9)+d*x(10)+e*x(11);...
    a*x(8)+b*x(9)+c7*x(10)+d*x(11)+e*x(12);...
    a*x(9)+b*x(10)+c8*x(11)+d*x(12)+e*x(13);...
    a*x(10)+b*x(11)+c9*x(12)+d*x(13)+e*x(14);...
    a*x(11)+b*x(12)+c10*x(13)+d*x(14)+e*x(15);...
    x(14)-2*x(13)+x(12);...
    R/(2*h^3)*(x(11)-2*x(12)+2*x(14)-x(15))]
end

Answer 1

[Umeshraja]

Hi @汶俊,

The objective function supplied to 'fsolve' must return a vector without any Inf or NaN entries.

Please refer to the following documentation on 'fsolve' function

http://www.mathworks.com/help/optim/ug/fsolve.html

Hope this helps!

Answer 2

[Torsten]

In the computation of c0,...,c10, you divide by x(3),...,x(13), thus by 0 because x0 = [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] .

After changing x0, you will most probably get complex values for the x-vector because some of the x become negative and you use x(i).^0.333 for 3<=i<=13 in your code.