I am trying to compute an integral using quad2d. The function that I'm integrating is also computed using quad2d. I suspect this might be the problem. I get the following error message
??? Error using ==> quad2d at 124 D must be a finite, scalar, floating point constant or a function handle.
Error in ==> H at 5 F=quad2d(@(x,z) fxw(x,z,lam,mu,sig),lb,ub,lb,w);
Error in ==> @(x,w)fxw(x,w,lam,mu,sig).*H(w,mu,lam,sig,V).*w
Error in ==> quad2d>tensor at 355 Z = FUN(X,Y); NFE = NFE + 1;
Error in ==> quad2d at 169 [Qsub,esub] = tensor(thetaL,thetaR,phiB,phiT);
Error in ==> QtleMax at 94 axm = quad2d(@(x,w) fxw(x,w,lam,mu,sig).*H(w,mu,lam,sig,V).*w , lb, ub, lb, ub)
Here is my code
close all;
clc;
clear all;
lam=[0.5 0.2 0.3];
mu=[-2 1 2];
sig=[1 0.5 1.5];
avmu=lam*mu';
mu = mu -avmu;
lb = min(mu)-6*max(max(sig),2);
ub = max(mu)+6*max(max(sig),2);
V = 38;
axm = quad2d(@(x,w) fxw(x,w,lam,mu,sig).*Hx(V,Fw(w,mu,lam,sig)).*w , lb, ub, lb, ub)
function y = fx(x,lam,mu,sig)
k1=max(size(lam));
k2=max(size(mu));
k3=max(size(sig));
if k1==k2 && k2==k3
y=0;
for i=1:k1
y=y + lam(i)*normpdf(x,mu(i),sig(i));
end
end
function y = fxw(x,w,lam,mu,sig)
y=normpdf(w-x,0,1).*fx(x,lam,mu,sig);
end
function y=Fw(w,lam,mu,sig)
lb = min(mu)-6*max(max(sig),2);
ub = max(mu)+6*max(max(sig),2);
y=quad2d(@(x,z) fxw(x,z,lam,mu,sig),lb,ub,lb,w);
end
function pr = Hx(V,F)
odd=mod(V,2);
if odd==1
pr= exp(sum(log(1:1:V-1)) - 2*sum(log(1:1:(V-1)/2)) + ((V-1)/2).*(log(F)+log(1-F)));
else
pr =exp(sum(log(1:1:V-1)) - sum(log(1:1:(V/2-1))) - sum(log(1:1:(V/2)))-log(2) + (V/2-1).*(log(F) + log(1-F)));
end
Please help me to figure out what the problem is
Thanks in advance, B.
