1
$\begingroup$

More generally, with $y=f(x)$ as above, can $f(A+B)$ be expressed in terms of $f(A)$ and $f(B)$? Other things I'd like are a Laurent series for f and, if it were possible, a linear differential equation for f with polynomial coefficients and at most polynomial forcing function. I don't know much about differential equations, but I suspect that last request is impossible. Thanks!

  • 2
    Try to find $x_1,x_2$ with $f(x_1)=f(x_2)$ but $f(2x_1)\neq f(2x_2)$. Pretty sure this is possible.2017-02-16
  • 0
    In particular, there are lots of values $x$ where $f(x)=0$...2017-02-16
  • 0
    Thomas, I think I get your clever idea in your first comment, but I didn’t understand the second comment: If f(x)=O=f(w), then doesn’t f(2x) =f(2w)=O?2018-10-15

1 Answers 1

2

In the case where $A=B=x$ $$f(2x) = \frac{\cos^2(x) - \sin^2(x)}{\sinh^2(x)+\cosh^2(x)}=\frac{\cos^2(x)-1/2}{\cosh^2(x)-1/2}$$ We note that this is fairly similar to being $f(x)^2$, and thus $f(x)^2$ is a fantastic approximation for any decently large $x$. We could even split the numerator to get closer to the form the OP desires; however, I am not confident that a simple closed form in terms of $f(x)$ exists

  • 0
    This is an interesting observation, but it doesn't answer the question.2017-02-16
  • 0
    @MathematicsStudent1122 True. More of a long comment. Nevertheless, I thought it interesting enough to share.2017-02-16