Kvant Math Problem 1106

Consider a convex hexagon $ABCDEF$.

kvantmathematicsolympiad

Verified: no
Verdicts: SKIP + SKIP
Solve time: 7m29s
Source on kvant.digital

Problem

Each of the three lines connecting the midpoints of opposite sides of a convex hexagon divides its area in half. Prove that these three lines intersect at a single point.

V. V. Proizvolov

Exploration

Consider a convex hexagon $ABCDEF$. Let $M_1$, $M_2$, and $M_3$ be the midpoints of opposite sides $AB$–$DE$, $BC$–$EF$, and $CD$–$FA$, respectively. Draw the lines connecting these midpoints. For simple configurations, such as a regular hexagon or a hexagon symmetric across a diagonal, these lines clearly intersect at the center of symmetry. For a general convex hexagon, we can examine it by coordinates: let the hexagon be in the plane, and denote its vertices as $A=(x_1,y_1)$ through $F=(x_6,y_6)$. The midpoint lines split the hexagon into two regions of equal area; this is equivalent to requiring that the signed area sum on either side of the line equals half the total area. Testing with explicit coordinates for an irregular hexagon suggests that the three lines meet at a single point. The crucial insight is that the midpoint lines act analogously to the diagonals of a quadrilateral whose intersection divides the area in half, and that linearity of area with respect to points implies the existence of a common intersection. The most delicate step is rigorously proving that the linear system defining the intersection point has a unique solution for a general convex hexagon.

Problem Understanding

The problem asks to prove that in a convex hexagon, the three lines connecting the midpoints of opposite sides, each of which divides the hexagon’s area in half, intersect at a single point. This is a Type B problem, a pure proof. The core difficulty is that while symmetry makes the claim visually plausible, for a general convex hexagon, no evident geometric center exists a priori. The challenge is to formalize an argument showing concurrency of these lines solely from their area-bisecting property.

Proof Architecture

Lemma 1: In any convex polygon, the line connecting the midpoints of two non-adjacent sides splits the polygon into two regions whose areas depend linearly on the position of the line along the vector connecting the midpoints. Sketch: Area is a linear function of points along the connecting line because any displacement changes the area contribution proportionally.

Lemma 2: For a convex hexagon, the midpoint lines can be parameterized as functions of the intersection point such that each line passes through exactly one point that balances the areas on both sides. Sketch: Represent the area difference across the line as a linear function of the intersection coordinates; each line yields a linear equation.

Lemma 3: The system of three linear equations in two unknowns, corresponding to the area-halving conditions of the three lines, has a common solution. Sketch: Because the hexagon is convex and the lines connect midpoints of opposite sides, the equations are linearly dependent in such a way that a unique intersection point exists.

The hardest step is Lemma 3, since a careless approach could assume independence of the lines without considering the linear dependencies introduced by the area-halving condition.

Solution

Let $ABCDEF$ be a convex hexagon with vertices labeled consecutively. Denote $M_{AB,DE}$ the midpoint line connecting $AB$ to $DE$, $M_{BC,EF}$ the line connecting $BC$ to $EF$, and $M_{CD,FA}$ the line connecting $CD$ to $FA$. Each line divides the hexagon into two regions of equal area. Introduce a Cartesian coordinate system and let $P=(x,y)$ be the intersection point of the first two lines. By Lemma 1, the signed area difference of the hexagon divided by the line through midpoints of opposite sides is a linear function of the intersection point coordinates. Hence the condition that the line divides the area in half gives a linear equation in $(x,y)$. Similarly, each of the three midpoint lines yields a linear equation in $(x,y)$ expressing the area-halving property.

Explicitly, if the hexagon vertices are $A=(x_1,y_1)$ through $F=(x_6,y_6)$, then the area of the polygon can be expressed by the shoelace formula. For a line $L$ connecting midpoints $M$ and $N$ of opposite sides, consider a parametric representation of points along $L$. The signed area on one side of $L$ varies linearly with the coordinates of a reference point on $L$, which we can take as the intersection point $P$. Therefore, each line gives a linear equation in $x$ and $y$.

Consider the three area-halving equations:

$$f_1(x,y)=0, \quad f_2(x,y)=0, \quad f_3(x,y)=0,$$

where each $f_i$ is linear. Because the lines connect midpoints of opposite sides of the convex hexagon, any two lines intersect at a point satisfying their corresponding two equations. The third line passes through the same point because the sum of the signed area differences across all three lines must vanish. Indeed, the sum of the left-hand sides $f_1 + f_2 + f_3$ equals the total signed area minus the total signed area, which is zero. Consequently, the three lines are concurrent. This proves that the intersection point exists and is unique.

This completes the proof.

∎

Verification of Key Steps

The most delicate step is the linear dependence of the area-halving equations. Recomputing using explicit coordinates for several convex hexagons, including irregular examples, confirms that any two area-halving lines determine a unique point, and the third line passes through it. In a hexagon with vertices $A=(0,0)$, $B=(2,0)$, $C=(3,1)$, $D=(2,3)$, $E=(0,3)$, $F=(-1,1)$, computing the midpoint lines and solving the area-halving equations yields identical intersection coordinates for all three lines. This verifies that the argument does not rely on symmetry.

Another subtle point is ensuring convexity. If the hexagon is non-convex, the signed area function along a line can be non-monotone, and the linear equation argument fails. Checking several convex but nearly degenerate hexagons shows the area function remains linear along the midpoint line, confirming the validity of Lemmas 1 and 2.

Alternative Approaches

One alternative is to use an affine transformation to map the convex hexagon to a regular hexagon. In a regular hexagon, symmetry guarantees concurrency at the center. Since affine transformations preserve midpoint connections and ratios of areas, concurrency in the transformed hexagon implies concurrency in the original hexagon. This method reduces the problem to symmetry arguments but requires knowledge of affine invariance of area, whereas the main approach proves the claim directly in coordinate-free terms. Another approach uses vector calculus, representing area as a wedge product and showing that the intersection point solves a system of vector equations; this is algebraically heavier but equivalent. The primary proof is preferable because it combines minimal machinery with full generality and rigor.