I've seen many times the terms like "cusp" and "corner" in the calculus books but I want a formal and rigorous definition for it . I know $f(x) = |x^2 - 1|$ has a cusp at $x = 1$ but I'm in doubt about $g(x) = \sqrt[3]{x^2}$ at $x= 0$ .
Definition of cusp or corner
1.7k Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail AtThere are 2 best solutions below
On
The example that you give, of the curve $y^3=x^2$, corresponds to what is called a regular cusp or ordinary cusp, in the sense that it is the least degenerate from the point of view of singularity theory. This leads us to:
Definition 1 (used by some authors): a cusp is precisely a point where a curve is equivalent to $y^3=x^2$ via a smooth local change of variables.
Definition 2 (more general): A piecewise smooth curve in, say, a euclidean space: $$f: \mathbb{R}\to\mathbb{R}^n$$ has a cusp at a point $p=f(x)$ if there is $\varepsilon>0$ such that $f$ is differentiable on both intervals $(x-\varepsilon, x)$ and $(x, x+\varepsilon)$, and if $$\lim_{h\to 0^+} f_1'(x-h)=\lim_{h\to 0^+} f_1'(x+h)$$ where $f_1'$ denotes the unit tangent vector to the curve.
In words, the curve leaves the point in the same direction as it has reached it.
At an elementary stage, when one talks about a cusp or a corner of a curve $h(t)$ at $x_0$, it would mean that:
$\lim_{t\rightarrow 0}\frac{h(x_0+t)-h(x_o)}{t} $ doesn't exist and in particular it is different when $t$ approaches $0$ from the left and from the right.
It is customary to talk of cusps to curves that are otherwise smooth (infinitely differentiable) and while continuous at $x_0$ they have a singularity.
As Arnaud Mortier noted in the comments, in more sophisticated settings (e.g. algebraic or differential geometry) cusp has a more complex and precise definition.