Calling a function normally vs feval

I have a script containing equations getting passed to lsqnonlin. Some of the equations need to call functions while they are being run like for e.g

eq(1) = x(1)^2 + 1 + f1(x(1))

eq(2) = 2x(1)^3 + x(2) + f2(x(2),x(1))

where x(1),x(2) are the variables.

I am using feval and using the function handle and passing the parameters, it's working all right. Why can't I use the normal function call like below? My exact code is below. The 2nd one is giving wrong answers. Am I missing something? P.S this isn't the code for the above equation.

eq(1) = K_1*((feval(@act1,x(2), x(6)))*(feval(@act2,x(2), x(3)))^2) - (feval(@act3,x(4),x(8),T)); %feval
%vs
eq(1) = K_2*(act2(x(2), x(6)))*(act5(x(2), x(3)))^2 - (act4(x(4),x(8),T)); %function call

Answer 1

[Walter Roberson]

eq(1) = K_1*act1(x(2), x(6)) .* act2(x(2), x(3)).^2 - act3(x(4),x(8),T);

The difference between feval() and calling the function directly, is that feval() permits passing in a character vector that is the name of a function to invoke. So for example instead of the user passing in @sin the user might pass in 'sin'

Using function handles instead of character vectors is almost always preferred, but there are a couple of routines such as arrayfun() that have special code for some common functions that permits faster execution than if a function handle were used.

[Angshuman Podder]

Thanks for your reply. Can you please tell why the 2nd formulation is giving wrong results?

[Walter Roberson]

act1, act2, act3 in one case. act2, act4, act5 in the other case. Different functions being called.