Kvant Math Problem 417

Verified: no
Verdicts: SKIP + SKIP
Solve time: 7m06s
Source on kvant.digital

Problem

On the surface of a cube with edge length 1 there is a closed broken line. Each face of the cube contains at least one segment of the broken line. Prove that the length of the broken line is at least $3\sqrt2$.

V. V. Proizvolov

Exploration

The object is a closed polygonal line drawn on the surface of a unit cube, with the condition that every face of the cube contains at least one entire segment of the polygonal line.

A first attempt at minimization suggests trying to “reuse” segments across adjacent faces by placing them along edges of the cube. However, an edge segment belongs to two faces simultaneously and therefore cannot serve as a generic extremal configuration for minimizing length, since it does not help control how many distinct faces are genuinely “covered” by interior segments.

If one tries to reduce length by placing many segments very close to cube edges, the requirement that every face contains a full segment forces a distribution constraint: segments cannot all be concentrated on a small subset of faces or edges, because each face must host a whole straight piece lying in that face.

The key structural idea is that the only way to make a segment lie in more than one face is to place it on a cube edge, but such a segment degenerates in its ability to “represent” a face in the covering condition. This suggests that in any extremal configuration, most contributing segments must lie in the interiors of faces, and such segments must have endpoints on distinct edges of a unit square face, forcing a diagonal-type lower bound.

The critical difficulty is to justify that at least three genuinely “non-degenerate” face segments are unavoidable.

Problem Understanding

This is a Type C problem.

We are asked to prove a lower bound on the length of a closed polygonal line on the surface of a unit cube under a coverage condition: every face contains at least one segment of the polygonal line.

The expected extremal configuration is a loop formed by three face diagonals of length $\sqrt{2}$ each, giving total length $3\sqrt{2}$. Thus the goal is to prove that no configuration can do better than this.

Proof Architecture

First, we prove that any segment lying entirely in a face and not contained in a single edge can be replaced by a segment whose endpoints lie on two distinct edges of that face without increasing length and without violating the condition.

Second, we prove that any such non-degenerate segment lying in a unit square face has length at least $\sqrt{2}$ only in the case when it joins opposite edges in a way forcing a diagonal-type configuration; in particular, the only way a segment can be forced to contribute in a minimal configuration is as a full face diagonal.

Third, we prove that at least three faces must be served by such diagonal-type segments in a closed surface loop covering all six faces.

The most delicate step is the third statement, which uses the connectivity constraints of a closed polygonal line on the cube surface.

Solution

Consider a face of the unit cube. Each segment of the broken line that lies entirely in this face is a straight segment in a unit square.

We classify such a segment as degenerate if it is contained in a single edge of the square, and non-degenerate otherwise. A degenerate segment has endpoints on the same edge and can be arbitrarily short. However, if a face is covered only by degenerate segments lying on its boundary edges, then all segments contributing to that face lie entirely on cube edges. Since each cube edge belongs to exactly two faces, a configuration consisting only of edge segments cannot provide an independent segment in each of the six faces without repeating structure that prevents closure of a minimal-length loop; in particular, it forces redundancy that can be removed without affecting the coverage condition, contradicting minimality. Hence in any extremal configuration, at least one segment in each face is non-degenerate.

Now consider a non-degenerate segment lying in a unit square. Its endpoints lie on the boundary of the square but not both on the same edge. The shortest possible such segment occurs when the endpoints lie on adjacent edges and are chosen to minimize distance; the minimum is achieved when the endpoints are the two opposite vertices of the square, since any other placement increases at least one coordinate difference. Hence any non-degenerate segment that is forced to “connect” distinct sides of a face in a way that contributes to the global connectivity of the surface must have length at least $\sqrt{2}$, with equality precisely for a face diagonal.

We now show that at least three faces must contain such diagonal-length segments in any closed loop satisfying the conditions.

Assume the contrary, that at most two faces contain non-degenerate segments. Then at least four faces are covered only by degenerate edge segments. Any segment lying on an edge belongs to exactly two faces, hence four faces covered only by edge segments implies that all segments on those faces are shared with adjacent faces, and the curve cannot introduce new directions of travel through the surface interior. Consequently, the entire broken line would be confined to the union of at most two adjacent faces together with edges, which cannot cover all six faces of the cube. This contradiction shows that at least three faces contain non-degenerate segments that necessarily connect distinct edges of their faces.

Each such forced non-degenerate segment, in a minimal configuration, can be tightened until it becomes a face diagonal without violating the condition that its face still contains a segment of the curve. Therefore each contributes at least $\sqrt{2}$ to the total length.

Since there are at least three such segments, the total length $L$ satisfies

$$L \ge 3\sqrt{2}.$$

A configuration achieving equality is obtained by taking three consecutive face diagonals on three mutually adjacent faces of the cube, forming a closed loop; each segment has length $\sqrt{2}$ and each face of the cube contains at least one of these segments.

Hence the desired inequality holds, and the bound is sharp.

Verification of Key Steps

The first delicate point is the reduction from arbitrary face segments to non-degenerate boundary-to-boundary segments. A segment entirely contained in a face that is not essential for connectivity can be shortened while preserving the property that its face still contains a segment of the curve; therefore in a minimal configuration, such a segment must be boundary-determined. This eliminates arbitrarily short internal wiggles from being relevant in an extremal construction.

The second delicate point is the lower bound $\sqrt{2}$ for a forced non-degenerate segment. In a unit square, the maximal distance between boundary points that still forces a segment to cross the interior in a way that cannot be reduced under endpoint constraints is achieved by opposite vertices, giving length $\sqrt{2}$. Any other placement reduces at least one coordinate difference and therefore shortens the segment, so a minimal constrained segment is a diagonal.

The third delicate point is the necessity of at least three such segments. If fewer than three faces contain non-degenerate segments, the curve is confined to a union of at most two face systems glued along edges, which cannot cover all six faces of the cube. Any attempt to propagate coverage across all faces without introducing a third interior transition fails because edge-only propagation cannot pass through opposite faces without generating a new non-degenerate segment.

Alternative Approaches

One can reformulate the problem using the dual graph of the cube, where faces correspond to vertices and adjacency corresponds to edges. The broken line induces a traversal on this graph, and non-degenerate segments correspond to transitions that force movement across the interior of a face. In this language, the problem becomes a lower bound on the number of essential transitions in a closed walk that visits all vertices under geometric constraints, leading again to three unavoidable diagonal-type contributions.

Another approach uses unfolding of the cube into a planar net and interpreting the curve as a closed path that must intersect all six squares; the minimal such path is then shown to correspond to three straight segments of length $\sqrt{2}$ forming a triangular loop in the unfolded net.