1

the question is 9.f ) section 5.4.6 from Mathematical Analysis I by Zorich page 263:

Choose the constants a, b, and R so that the circle $(x − a)^2 + (y − b)^2 = R^2$ has the highest possible order of contact with the given parametrically defined curve $x = x(t)$, $y = y(t)$. It is assumed that $x(t)$ and $y(t)$ are twice differentiable and that $(\dot x(t_0),\dot y(t_0)) \neq (0, 0)$. This circle is called the osculating circle of the curve at the point $(x_0, y_0)$....ect.

I went back to page 181 and here is the definition for "order of contact":

Definition 5: If the mappings $f : E \longrightarrow R$ and $g : E \longrightarrow R$ are continuous at a point $x_0 ∈ E$ that is a limit point of $E$ and $f (x) − g(x) = o((x − x_0 )^n )$ as $x \longrightarrow x_0$ , $x ∈ E$, we say that $f$ and $g$ have $n^{th}$ order contact at $x_0$ (more precisely, contact of order at least $n$).
For $n = 1$ we say that the mappings $f$ and $g$ are tangent to each other at $x_0$ .

How do I even start? like is there a formula of Taylor expansion for implicit functions introduced before ?

zaknenou
  • 169

2 Answers2

1

I do not have the book by Zorich but here you have to consider the real variable function $t\to f(x(t),y(t))$ where $f(x,y)=(x-a)^2+(y-b)^2-R^2$ and the constant $0$ function and ask which is the highest possible contact order.

In other words you Taylor expand $t\to f(x(t),y(t))$ and start annihilating terms in its expansion.

  • I did what you said, calculated the first two derivatives of $f(t)$ and set them all equal to zero thus getting 3 equations for $a, b$ and $R^2$ and I got $R^2=\frac{(\dot x^2+\dot y^2)^3}{(\ddot y\dot x-\ddot x\dot y)^2}$ which looks like the right answer. I'll write a full answer later and credit you (really thanks for the help). But can you explain why this ? I mean from the definition the forward approach is to use the parametric equations of circle and curve, calculate the distance between $D(t)=||\gamma(t)-C(t)||$ and do Taylor expansion or derivatives, why does this work too ? – zaknenou Dec 14 '24 at 21:25
  • 1
    If I've understood correctly your question: if you have two parametric curves then you just compare parametric expansions, but if one of them is given as a regular level set F(x,y)=0 then you can Taylor expand $F\circ \gamma$. To prove that the two approaches are in fact equivalent requires a proof (basically just based on chain rule); at present I do not have a reference at hand for that – N. Ciccoli Dec 16 '24 at 09:20
0

let $\text{C} : \mathbb{R} \to \mathbb{R^2}, \quad t \mapsto (x_C(t),y_C(t))$ be the parametrization of our circle. Then: $$\begin{pmatrix} x_C(t) \\ y_C(t) \end{pmatrix}=\begin{pmatrix} a + R\cos t \\ b + R\sin t \end{pmatrix}$$ Define also the function $f : \mathbb{R^2} \to \mathbb{R}, \quad (x,y) \mapsto (x(t) − a)^2 + (y(t) − b)^2 - R^2$, so that the Circle $C$ is also defined by equation $f(x,y)=0$.

following @N.Ciccoli : $C$ and $\gamma$ have $n^{th}$ order of contact at $t_0$ means that $(f\circ \gamma)(t)$ is $o((t − t_0 )^n )$ as $t \longrightarrow t_0$.
from my understanding, reason is: since $x(t)$ and $y(t)$ are both assumed twice differentiable and not vanishing together:

  • if $C$ and $\gamma$ do have order of contact $0$ then $\gamma(t_0)=\text{C}(t_0)$, which means that $\gamma(t_0)=(x(t_0),y(t_0))=(x_0,y_0)$ satisfies the implicit equation of $C$, $$(x_0 − a)^2 + (y_0 − b)^2 - R^2=0 \tag{1}$$
  • if they have order of contact $1$ then $\gamma'(t_0)=\text{C}'(t_0)=(\dot x_C(t_0),\dot y_C(t_0))$. But the implicit function theorem tells us that $2(x_C-a)\dot x_C+2(y_C-b)\dot y_C=0$, $\forall t \in \mathbb{R}$ and thus $2(x_0-a)\dot x_0+2(y_0-b)\dot y_0=0\implies$ $$(x_0-a)=-\frac{\dot y_0}{\dot x_0}(y_0-b) \tag{2}$$
  • same reasoning for when they do have order of contact 2 yields: $2\ddot x_0(x_0-a)+2\dot x_0^2+2\ddot y_0(y_0-b)+2\dot y_0^2=0$ substituting $(x_0-a)$ using the previous equation (2): $$(y_0-b)=-\frac{\dot x_0^2+\dot y_0^2}{\ddot y_0-\frac{\ddot x_0}{\dot x_0} \dot y_0} \tag{3}$$

...ect, the same reasoning goes for any order of contact if we assume $\gamma$ to be smooth.
Now substituting both $(x_0-a)$ and $(y_0-b)$ in equation $(1)$ so that we get an equation with only R and derivatives of $x_0(t)$ and $y_0(t)$, first substituting $(x_0-a)$:
ps: I'll be omitting the variable here, so $\dot x$, $\dot y$, $\ddot x$, $\ddot y$ are all evaluated at $t_0$ in these calculations. Just abbreviation here.

$$\left(\frac{\dot y}{\dot x}\right)^2(y-b)^2+(y-b)^2=R^2$$ i.e $$(y-b)^2\left(\frac{\dot y^2}{\dot x^2}+1\right)=R^2\implies (y-b)^2 \times \frac{\dot y^2+\dot x^2}{\dot x^2}=R^2$$
Now substituting $(y-b)$: $$\frac{\left(\dot x^2+\dot y^2\right)^2}{\left(\ddot y-\frac{\ddot x}{\dot x} \dot y\right)^2} \times \frac{\left(\dot x^2+\dot y^2\right)}{\dot x^2}=R^2$$ i.e $$R^2=\frac{\left(\dot x^2+\dot y^2\right)^3}{\left[\dot x\left(\ddot y-\frac{\ddot x}{\dot x} \dot y\right)\right]^2}=\frac{\left(\dot x^2+\dot y^2\right)^3}{\left(\dot x\ddot y-\ddot x \dot y\right)^2}$$
Again $\dot x$, $\dot y$, $\ddot x$, $\ddot y$ are all evaluated at $t_0$, they were only abbreviated. Thus: $$R_0=\frac{\left(\dot x_0^2+\dot y_0^2\right)^\left(\frac{3}{2}\right)}{\left|\dot x_0\ddot y_0-\ddot x_0 \dot y_0\right|}$$ Because R is dependent on $t_0$ too. From that we conclude $a$ and $b$: $$\begin{pmatrix} a \\ b \end{pmatrix}=\begin{pmatrix} x(t_0) - R(t_0)\cos t_0 \\ y(t_0) - R(t_0)\sin t_0 \end{pmatrix}$$ Since a, b and R are already determined there is no going further to order of contact $3$ I guess, the solution ends here, I think.

zaknenou
  • 169