B Is there a function for every possible "path"?

  • B
  • Thread starter Thread starter Jurgen M
  • Start date Start date
  • Tags Tags
    Function Path
Jurgen M
Has any path/line/shape/contour function? And how find function of complex "path"?

for example this?
graph.png
 
Mathematics news on Phys.org
Well since your function backtracks on itself then no there is no equation f(x) that can represent it.

Do you know why?

However, if you parametrize both x and y with some variable called t then maybe you can piece together a vector function to imitate the path above.
 
  • Like
Likes russ_watters and FactChecker
A path is best parameterized. You can parameterize the above path in the form of ##P: (x(t),y(t)), 0\le t \le1##.
 
This isn't a function in the sense of ##f(x)=y##. But you can "walk" along the path as a function ##[0,1]\longrightarrow \mathbb{R}^2.## In this case, you have to glue all partial paths to a single long one.
 
jedishrfu said:
Well since your function backtracks on itself then no there is no equation f(x) that can represent it.

Do you know why?

However, if you parametrize both x and y with some variable called t then maybe you can piece together a vector function to imitate the path above.
No.Why?

What is procedure to find f(x) of some given complex path that is not backtracks itslef?
 
Jurgen M said:
What is procedure to find f(x) of some given complex path that is not backtracks itslef?
You have to distinguish between the path, and the relation ##\{(x,f(x))\}##. What do you want to talk about? One is a function ##\mathbb{R} \longrightarrow \mathbb{R}^2## and the other a subset of ##\mathbb{R}^2##.

As long as this isn't answered, this thread will be a mess.
 
Jurgen M said:
No.Why?

What is procedure to find f(x) of some given complex path that is not backtracks itslef?
It can not be in the form of f(x) because, for a single value of x, there are multiple values of y that are matched with it on the path. Which y-value should f(x) be?
On the other hand, suppose you parameterize the path in the form of ##(x(t), y(t)), 0 \lt t \lt 1##. For any value of ##t \in [0,1]## you have exactly one (x,y) value.
 
  • Like
Likes russ_watters and PeroK
fresh_42 said:
You have to distinguish between the path, and the relation ##\{(x,f(x))\}##. What do you want to talk about? One is a function ##\mathbb{R} \longrightarrow \mathbb{R}^2## and the other a subset of ##\mathbb{R}^2##.

As long as this isn't answered, this thread will be a mess.
Slowly.
First I don't understand this math terms..

What is difference between "path" and f(x)?
 
For any value x there is one and only one y value. When the graph backtracks as yours does then for a given x there are more than one y values.

https://en.wikipedia.org/wiki/Function_(mathematics)

thats why you have to think of it more like the flight of a bumblebee whizzing around in 2D or 3D space and using the time t as the independent variable otherwise you'll get stung.
 
  • #10
Jurgen M said:
Slowly.
First I don't understand this math terms..

What is difference between "path" and f(x)?
I answered this in post #4. A path is a walkthrough of the line you drew (every point has a unique timestamp), from a starting point to an endpoint. If you consider ##f(x)## as a function graph, then it is a) not a function because there are points ##x## with two different values ##f(x)##, and b) a subset of the plane. Such a subset can be considered as a relation. Relations allow ambiguities, functions do not.
 
  • #11
It's a relation, as far as I know (I am speaking without background; but I find interesting this thread) You are never express algebraically that path without a huge effort.
 
  • #12
mcastillo356 said:
It's a relation, as far as I know (I am speaking without background; but I find interesting this thread) You are never express algebraically that path without a huge effort.
That's a guess. Therefore post #6. As long as everybody chooses for himself what the OP might have meant, as long will this thread be doomed to confusion.
 
  • Like
  • Informative
Likes russ_watters, mcastillo356 and jedishrfu
  • #13
mcastillo356 said:
It's a relation, as far as I know (I am speaking without background; but I find interesting this thread) You are never express algebraically that path without a huge effort.
You might be surprised at how well Fourier series can approximate such things. See this (especially the first few minutes).
 
  • Informative
Likes mcastillo356
  • #14
mcastillo356 said:
It's a relation, as far as I know (I am speaking without background; but I find interesting this thread) You are never express algebraically that path without a huge effort.
We distinguish between a "formula" and a "function".

A "function" is a mapping from elements in a "domain" to elements in a "range" such that each element in the domain is mapped to exactly one element in the range.

A "formula" like ##\sin x## or ##\sqrt{t}\ \hat{i} + t^2\ \hat{j}## can let you specify a function without having to write down a complete table for the mapping.

My guess (*nods to @fresh_42*) is that you are asking whether for every path you can draw with an ideal pencil on an ideal piece of paper there is a finite formula that describes that path with perfect fidelity.

The answer to that question is no. Unless I am misremembering, the cardinality of the set of continuous ideal paths is the cardinality of the continuum and the cardinality of the set of formulas is aleph null.
 
  • Informative
Likes mcastillo356
  • #15
jbriggs444 said:
The answer to that question is no. Unless I am misremembering, the cardinality of the set of continuous ideal paths is the cardinality of the continuum and the cardinality of the set of formulas is aleph null.
If only a picture is given, then we can get arbitrary close with formulas and piecewise-defined paths.

This is the next problem with the question: How is that scribbling given? How accurately do we have to be?
 
  • Like
Likes jbriggs444
  • #16
Just to be clear, when a path is parameterized by a variable ##t \in [0,1]##, like ##P: f(t) = ( x(t), y(t) ), t \in [0,1]##, that is a function from ##[0,1]## to ##\mathbb{R}^2##.
 
  • #17
FactChecker said:
Just to be clear, when a path is parameterized by a variable ##t \in [0,1]##, like ##P: f(t) = ( x(t), y(t) ), t \in [0,1]##, that is a function from ##[0,1]## to ##\mathbb{R}^2##.
Yes, and ##G:=\{(x,y)\in \mathbb{R}^2\,|\,y=f(x)\}\subseteq \mathbb{R}^2## is a relation, and in the case of ##f## as in post #1 not a function.
 
  • Like
Likes FactChecker
  • #18
I think this question is equivalent to whether every path can be parameterized. It seems that if it's rectifiable, it may be parameterized by arc length.
 
  • Like
Likes FactChecker
  • #19
WWGD said:
I think this question is equivalent to whether every path can be parameterized. It seems that if it's rectifiable, it may be parameterized by arc length.
It is something akin to cheating, but even if the path cannot be parameterized as a continuous function over the reals, one can nonetheless unambiguously associate a real number t with each (x,y) coordinate in the graph (e.g. by interleaving the digits in the decimal expansions) and then declare f(t) = (x,y) and set the domain to the set of t values that are used. It is unlikely that this function will be continuous, but it will be a function from [a subset of] ##\ \mathbb{R}\ ## to the set of coordinates of points with pencil marks.
 
Last edited:
  • Like
Likes WWGD
  • #20
jbriggs444 said:
It is something akin to cheating, but even if the path cannot be parameterized as a continuous function over the reals, one can nonetheless unambiguously associate a real number t with each (x,y) coordinate in the graph (e.g. by interleaving the digits in the decimal expansions) and then declare f(t) = (x,y) and set the domain to the set of t values that are used. It is unlikely that this function will be continuous, but it will be a function from [a subset of] ##\ \mathbb{R}\ ## to the set of coordinates of points with pencil marks.
How about just using the ortho projections to the respective axes as the coordinates. Wonder if this is specific enough for the op?
 
Back
Top