Integration error - "integrand F must return a single, real-valued vector of the same size as the input" - even when the integrand DOES return such a vector

I call the following within a function:

Vsame = @(x) Vsame_fit(x./lB);
F0011 = @(x) (1-(x.^2)/2).*exp(-(x.^2./2));
integ_F0011same = @(x) F0011(x).*Vsame(x);
integ_F0011same(linspace(1,3,10))
Fsame(1,1,2,2) = quadcc(integ_F0011same,0,Inf);

I defined the constant lB elsewhere. Outside that function I have defined Vsame_fit:

function y = Vsame_fit(q)
	for k=1:length(q)
		if q(k)<0.022913
			y(k) = 0.98*tanh(200*q(k));
		elseif q(k)<0.50274
			y(k) = 1/(0.9*q(k)+1);
		elseif q(k)<21.598
			y(k) = 1/(1.046*q(k)+0.9266);
		else
			y(k) = 1/(0.9512*q(k)+2.89);
		end
	end
end

I wrote the loop over the length of the argument to ensure that it would return a vector. Furthermore, if I call something simple as a test like

integ_F0011same(linspace(1,3,10))

inside the function, a 10-element is returned as expected. Why am I getting this error?

At the end of the day, I just want to have the value of the integral which I was hoping would be saved to Fsame(1,1,2,2). The reason I have defined things with so many separate functions, however, is that I reuse Vsame three more times to evaluate similar integrals, e.g.

integ_F0110same = @(x) F0110(x).*Vsame(x);
Fsame(1,2,2,1) = quadcc(integ_F0110same,0,Inf);

where F011(x) was also defined with a function handle. Is there a better way I should be doing this?

[Matt J]

Because quadcc is not a Mathworks-shipped function, you will probably have to ask its author.

[Brett Green]

Thanks. It actually does execute properly when I change quadcc to quad, but I thought quad was not recommended for use on nonsmooth integrands, of which Vsame is an example because it's defined piecewise.

Answer 1

[Steven Lord]

For what it's worth, when I defined your functions (with lB equal to 1, since you didn't specify a value) and called integral on it I received a result. Unless you're required to use quadcc, I'd just switch to integral.