3
$\begingroup$

The unit circle under the $L_2$ norm can be described parametrically by

$x = \cos(t), y = \sin(t), 0 \le t \le 2\pi$

How do we parametrically describe the unit circle in general $L_p$ space (for simplicity, assume $1 < p < \infty$)? Is it reasonable to view the appropriate functions $x = \cos_p(t), y = \sin_p(t)$ as generalizations of the usual definitions of $\sin, \cos$?

$\endgroup$

2 Answers 2

3
$\begingroup$

I suspect this might be a preference question, but Generalized spherical and simplicial coordinates [W.D.Richter 2007] Suggests that the correct unit circle identity would actually be:

$\mid\cos_p(t)\mid^p + \mid\sin_p(t)\mid^p = 1$

He defines the trig functions as:

$\cos_p(t) = \dfrac{\cos(t)}{N_p(t)}$ and $\sin_p(t) = \dfrac{\sin(t)}{N_p(t)}$ with $N_p(t)=(\mid\cos(t)\mid^p + \mid\sin(t)\mid^p)^{\frac{1}{p}}$

My sense is that his unit circle identity seems more analogous to other expressions involving p-norms than yours is.

$\endgroup$
1
$\begingroup$

[I think this was an easy question that I asked too quickly, and I can actually answer my own question.]

I think the right analog is $$\cos_p(t) := \cos^{2/p}(t), \sin_p(t) := \sin^{2/p}(t)$$

We then have $\cos_p(t)^p + \sin_p(t)^p = \cos(t)^2 + \sin(t)^2 = 1$, which means that we have correctly described the unit circle.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.