Finding Connected Sets in a Rectangle

  • Context: Graduate 
  • Thread starter Thread starter HallsofIvy
  • Start date Start date
  • Tags Tags
    Rectangle Sets
Click For Summary

Discussion Overview

The discussion revolves around the problem of finding two disjoint connected sets, P and Q, within a closed rectangle in R2. The sets must contain specific diametrically opposite points and satisfy certain conditions regarding their connectedness and disjointness. The conversation explores various proposed constructions and the nuances of connectedness in topology.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • One participant proposes that P consists of points where x is rational and y is irrational, while Q consists of points where x is irrational and y is rational, both sets containing the points (1, 1) and (-1, -1).
  • Another participant suggests a construction for P that includes two disjoint parts mirroring each other and a straight segment for Q, arguing that this approach leads to a topological structure similar to the topologist's sine curve.
  • A different construction is proposed where P is defined as a union of vertical and horizontal line segments, described as interleaved combs, and Q is its complement, with claims about the connectedness of these sets.
  • One participant challenges the connectedness of the proposed sets, suggesting that P can be disconnected by splitting it based on irrational values of x.
  • Another participant introduces a new example involving straight lines and sinusoidal functions to demonstrate connectedness, asserting that the closures of certain sets include necessary limit points.
  • Further discussion includes a mention of a simpler approach involving the sine function, indicating ongoing exploration of the problem.

Areas of Agreement / Disagreement

Participants express differing views on the connectedness of the proposed sets, with some asserting that their constructions are valid while others challenge these claims. The discussion remains unresolved, with multiple competing views on how to define and construct the sets P and Q.

Contextual Notes

Some constructions rely on specific properties of rational and irrational numbers, and the definitions of connectedness may vary based on the interpretations of the participants. There are unresolved mathematical steps regarding the connectedness of the proposed sets.

HallsofIvy
Science Advisor
Homework Helper
Messages
42,895
Reaction score
983
Since things are a bit quiet here, I thought I would throw out a puzzle I came up with several years ago, after reading an article on connected sets:

Find two sets, P and Q, satisfying:

1) Both P and Q are completely contained in the (closed) rectangle in R2 with vertices at (1, 1), (1, -1), (-1, -1), and (-1, 1).

2) P contains the diametrically opposite points (1, 1) and (-1, -1) while Q contains(1, -1) and (-1, 1).

3) P and Q are both connected sets.

4) P and Q are disjoint.

The solution involves the difference between "connected" and "path-wise connected".
 
Physics news on Phys.org
P is the set of points [itex](x,y)[/itex] such that [itex]x[/itex] is rational and [itex]-1\leq x\leq 1[/itex], and [itex]y[/itex] is irrational and [itex]-1\leq y \leq 1[/itex] along with (1,1) and (-1,-1).

Q is the set of points [itex](x,y)[/itex] such that [itex]x[/itex] is irrational and [itex]-1\leq x\leq 1[/itex], and [itex]y[/itex] is rational and [itex]-1\leq y \leq 1[/itex] along with (1,1) and (-1,-1).
 
Last edited:
AKG said:
Make Q the straight segment. Make P have two disjoint parts, each the mirror image of the other. Any half of P will start at its corner, and approach Q like the topologist's sine curve approaches the y-axis.

P can be partitioned into the closed sets that correspond to [itex]x+y > 0[/itex] and [itex]x+y < 0[/itex]. Since solutions to [itex]x+y=0[/itex] are not in P, both of those sets are closed (in P), and they're obviously disjoint.
 
Nate, your idea doesn't work. P and Q are not connected. You can disconnect P say be splitting it into two sets, those points in P to the left of the vertical line x=any irrational in (-1,1) and those to the right.
 
Editing to make things nicer...

Consider the following sets:
[tex]P[/itex]<br /> is a union of the following:<br /> A vertical line segment at [itex]x=-1, y \in (-1,.5][/itex]<br /> A vertical line segment at [itex]x=1, y \in [-.5,1)[/itex]<br /> Horizontal line segments [itex]x \in [-1,.5], y = \sqrt{2} * n - k (\rm{for some} k,n \in \mathbb{Z} \rm{and} y \in [-1,.5][/itex]<br /> and<br /> Horizontal line segments [itex]x \in [-.5,.-1], y = \sqrt{2} * n - k (\rm{for some} k,n \in \mathbb{Z} \rm{and} y \in [-.5,.1][/itex]<br /> <br /> This could be described as two interleaved infernal combs.<br /> And [itex]Q[/itex] is [itex]P[/itex]'s complement.<br /> Clearly each comb is a path-connected subset, so the only possible partition into non-empty closed sets is to split this into the combs, but each comb contains part of the other in its closure. Hence [itex]P[/itex] is connected.<br /> <br /> (I'm open to suggestions on how to improve this section.)<br /> Now, let's assume that [itex]Q[/itex] can be partitioned into disjoint non-empty closed sets A, and B. Since any horizontal intersecting [itex]Q[/itex] forms a connected sets, the projections of A and B onto the horizontal line must be disjoint. Since line segements are connected, at least one of two cannot be a closed set. Without loss of generality, that set is A. Since the projection of A onto the y-axis is not closed there is some [itex]y[/itex] that is a limit point of the projection, but not in the projection, but that limit point clearly corresponds to a line segment of limit points of A that are not in A, but are in [itex]A[/itex] - contradicting that A is closed.[/tex]
 
Last edited:
If anyone cares, I thought of a better example (in the sense that it's a bit easier to prove connectedness) on my way home last night, but it uses the same idea.
 
NateTG, I'm going to have to think about that for a while. Here' my answer:

Let A be the straight line from (-1,-1) to (0, -1/2). Let B be the set
{(x,y)| 0< x<= 1/pi, y= 0.8 sin(1/x)+ .1} (0.8 and 0.1 are chosen to lift that slightly above the x-axis but stay within the square), and let C be the straight line from (1/pi,0.1) to (1,1). Let P= A union B union C.

Let X be the straight line from (-1,1) to (0, 1/2). Let Y be the set {(x,y)| 0< x< 1/pi, y= 0.8 sin (1/x)- .1}, and let Z be the straight line from (1/pi,-0.1) to (1,-1). Let Q= X union Y union Z

Each of A, B, C, X, Y, Z is connected B union C and Y union Z are clearly connected since B,C and Y,Z have a point in common. The fact that
P= A union B union C is connected is clear from the fact that the closure of B includes the entire line (0, y) with y from -0.7 to 0.9 and includes (0, 0.5). The fact that Q= X union Y union Z is connected is clear from the fact that the closure of Y includes the entire line (0, y) with y from -0.9 to 0.7 and includes (0, 0.5).
 
HallsofIvy said:
NateTG, I'm going to have to think about that for a while. Here' my answer.

Nice.

I had thought about using
[tex](1-|x|) \sin (\frac{1}{x}) \pm {x}[/itex]<br /> but for whatever reason didn't think of simply splitting the y-axis into, for example, positive and negative sections.[/tex]
 

Similar threads

  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 80 ·
3
Replies
80
Views
10K
  • · Replies 16 ·
Replies
16
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 12 ·
Replies
12
Views
3K
  • · Replies 17 ·
Replies
17
Views
7K