Quad error, simple function,

I am a beginner, so please bear with me

cpn is referenced as such as this is part of a much larger code

clc

clear all

close all

cpn=[29105 8614.9 1701.6 103.47 909.79];

fun=@ (ll) cpn(1)+cpn(2)*((cpn(3)./ll)/sinh(cpn(3)./ll))^2+cpn(4)*((cpn(5)./ll)/cosh(cpn(5)./ll))^2

fun(100)

fun(200)

Q=quad(fun,18,200)

I am getting this error:

Error using quad (line 71) The integrand function must return an output vector of the same length as the input vector.

Error in Inttry (line 11) Q=quad(fun,18,200)

Thanks

Answer 1

[Roger Stafford]

Your function definition is lacking operators in at least four places, one after "cpn(2)", one after "cpn(4)", and one before each of the two 2's. As such it does not constitute a valid matlab function definition.

[adfnm]

Okay, so how should it look? I am very sorry for my ignorance, but I have no idea hoe to correct this. Also, the fun(100) and fun(200) do give valid answers when the code runs.

[adfnm]

Sorry, I do have those in the code

clc

clear all

close all

cpn=[29105 8614.9 1701.6 103.47 909.79];

fun=@ (ll) cpn(1)+cpn(2)*((cpn(3)./ll)/sinh(cpn(3)./ll))^2+cpn(4)*((cpn(5)./ll)/cosh(cpn(5)./ll))^2

fun(100)

fun(200)

quad(fun,100,2000)

[Roger Stafford]

That looks like something that would work properly. Does it work, and is it the function you intended as the integrand?

[adfnm]

Yes, that is the function. fun(100) returns a scalar, thats why I am baffled to the cause of this error

Answer 2

[Iain]

I guess that you've missed off declaring a few operations as elementwise operations, and it's getting out a square matrix instead of a vector.

Each divide, multiply and power should have a "." just before them.

[adfnm]

fun(100) returns a scalar, thats why I am baffled to the cause of this error, It is not the function that is the problem!

[Roger Stafford]

The documentation of 'quad' at

 http://www.mathworks.com/help/matlab/ref/quad.html

gives you the answer: "The function y = fun(x) should accept a vector argument x and return a vector result y, the integrand evaluated at each element of x."

This means that 'quad' doesn't send out just scalar quantities to be evaluated. It asks for entire vectors in x to be evaluated to the corresponding vectors y. That means that your expression which works for scalars but is inappropriate for vectors, will not succeed in 'quad'. You must put in those dots as Iain has recommended if you are to avoid the error messages and obtain a correct result.