4
$\begingroup$

The following problem seems a very hard one, is it known? It has a resemblance to the lonely runner conjecture. I am guessing.

In the plane let $v_i$ be $n$ unit vectors no two of them are colinear. Take $P_0$ any point in the plane and construct a successive set of points $P_i$ such that for every $i$, $1\le i\le n$ the segments $P_{i-1}P_{i}=v_j$ for some $j$. Every vector $v_j$ is used once. Can we choose the vectors so that the obtained path from $P_0\to P_n$ is either a simple cycle or has no crossings (planar)?

$\endgroup$
4
  • 1
    $\begingroup$ Do you mean you want to characterize the $n$-tuples of pairwise not collinear vectors for which you can make that choice? And if so, for every $n$? I don't really see the link with the lonely runner, by the way. $\endgroup$ Jul 30, 2021 at 16:20
  • $\begingroup$ Hey there is no link (just imagining) the $n$ vectors are given. You take a vector for each edge in the path... $\endgroup$
    – Toni Mhax
    Jul 30, 2021 at 17:08
  • 1
    $\begingroup$ So you are given $n$ unit vectors, and $n$ doesn't matter. You wonder whether you can always choose one out of $n$! in principle distinct permutations of your vectors so that the corresponding concatenation doesn't self-intersect. Is this your question? If so the answer is no: take 3 vectors parallel to the sides of an equilateral triangle oriented suitably. $\endgroup$ Jul 30, 2021 at 19:12
  • 1
    $\begingroup$ Crossing here is strictly speaking as X rather than V. Thanks $\endgroup$
    – Toni Mhax
    Jul 30, 2021 at 23:23

1 Answer 1

11
$\begingroup$

Let $S=\Sigma v_i$. If $S=0$, sort the vectors according to their angle along the unit circle. Then the corresponding closed path traces the boundary of a convex polygon.

In fact, the vectors $v_i$ can be of arbitrary length.

If $S\neq 0$, then add an auxiliary vector $v_{n+1}=-S$ and proceed as in the first case. Finally remove the segment given by the vector $v_{n+1}$.

$\endgroup$
3
  • $\begingroup$ Tx but you didn't follow the conditions the points need to be constructed successively in a one way directed path, here when you add a vector you add a point from nowhere.... $\endgroup$
    – Toni Mhax
    Jul 31, 2021 at 2:11
  • 2
    $\begingroup$ I don't understand your comment, Toni. Each new vector $v$ starts where the last one ended, and goes to wherever $v$ goes from there. Not adding any points from nowhere. It's just ordering the vectors in a clever way as $u_1,u_2,\dots$ and then going from the origin (say) to $u_1$ to $u_1+u_2$ and so on. $\endgroup$ Jul 31, 2021 at 6:03
  • 1
    $\begingroup$ Yes i see . Thank you $\endgroup$
    – Toni Mhax
    Jul 31, 2021 at 6:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.