2
$\begingroup$

We have a $n \times m$ table and for each cell we color one of it's diagonals prove that we have a path with colored edges from top of the grid to bottom of that or from left to right! Can anyone help me with solving this problem?

  • 0
    Can you clarify what you mean by the "diameter" of a cell? Are you referencing a single edge of the cell, all four edges of the cell, one of the two diagonals of the cell, etc?2017-02-28
  • 0
    I think he means diagonal.2017-02-28
  • 0
    edited sorry for mistake2017-02-28
  • 0
    Clearly true for 2*2. Can you use induction for 2*m then induction again for n*m?2017-02-28
  • 0
    I tried! But I don't think problem can be solved this way2017-02-28

1 Answers 1

3

Let $G$ be a graph, where all corners of cells are vertices and colored diagonals are edges. Also it has a left vertex $\ell$ adjacent to all leftmost corners and a right vertex $r$ adjacent to all rightmost corners.

Let $H$ be another graphs, where all middles of cell sides are vertices. Each cell gives to $H$ two internal edges that connect vertices of $H$ and don't cross colored diagonal. Also $H$ has a top vertex $t$ adjacent to all topmost middles and a bottom vertex $b$ adjacent to all bottommost middles.

If $G$ has a walk between $\ell$ and $r$ then it is what we need. It is not hard to see that otherwise $H$ has a walk between $t$ and $b$, that we can't overcome moving from $\ell$ to $r$. Inside each cell edges of $H$ are parallel to a colored diagonal. Now consider shortest walk $tv_1v_2\ldots v_kb$ between $t$ and $b$ and construct corresponding walk from top to bottom via colored diagonals. If $v_2$ is to the left of $v_1$ then we start with diagonal from the cell containing $v_1$ and $v_2$ on its sides. Otherwise $v_3$ is to the right of $v_2$ and we start with diagonal from the cell containing $v_2$ and $v_3$. Further for each pair of vertices $v_i$ and $v_{i + 1}$ of the walk we take diagonal from cell containing both vertices if it is adjacent or equal to previous diagonal taken or just skip otherwise. Selected diagonals give us walk from top to bottom. (Note that this way we can traverse some diagonals twice that is no problem, we can remove all such diagonals if needed.)

P. S. I guess there should be more elegant solution.

  • 0
    I really dont see how anything more elegant can exist. This reminded me of the proof that a compact surface of characteristic $2$ is a sphere.2017-02-28
  • 0
    that was a good solution tnx!2017-02-28
  • 0
    More elegant solution if you need: you can prove this with sperner lemma and induction2017-02-28
  • 0
    can you please explain how do you construct H and what is two edges for vertecies2017-02-28
  • 0
    [Here](http://imgur.com/a/gSWom) is an example of $G$ (red) and $H$ (green).2017-03-02