Kvant Math Problem 471

Verified: no
Verdicts: SKIP + SKIP
Solve time: 7m20s
Source on kvant.digital

Problem

Two intersecting circles divide the plane into three bounded disjoint regions. Prove that there does not exist a circle that bisects the area of each of these three regions.

S. V. Fomin

Exploration

Two intersecting circles partition the plane into exactly three bounded regions: the common lens $R_0$, the two asymmetric caps $R_1$ and $R_2$ lying respectively in the first and second circle but outside the other. The problem asserts that no circle can simultaneously bisect the area of each of these three regions.

A natural first reduction uses additivity of area. If a circle $\Gamma$ bisects each of $R_0,R_1,R_2$, then for each original circle $C_1,C_2$ it also bisects the total area inside them, since

$$C_1 = R_0 \cup R_1,\qquad C_2 = R_0 \cup R_2$$

with disjoint unions. Hence any such $\Gamma$ must bisect both $C_1$ and $C_2$.

Thus the problem reduces to proving that two intersecting circles do not admit a common area-bisecting circle.

The key geometric structure is that the intersection area of two circles depends only on the distance between their centers. This suggests parameterizing candidate bisecting circles by their center and radius, reducing each bisecting condition to a constraint on one distance variable. The difficulty is to show that the two resulting constraints are incompatible for a single pair of distances coming from a point in the plane.

The most fragile point is the transition from two area equations to a contradiction involving the geometry of three points, since one must ensure no hidden degeneracy allows simultaneous satisfaction.

Problem Understanding

This is a Type B problem: prove that no circle exists with a simultaneous area-bisecting property for three regions defined by two intersecting circles.

The core difficulty is global compatibility of two nonlinear area constraints coming from different circular intersections. The expected structure is that bisecting one circle already strongly restricts the position of the candidate center, and enforcing the second constraint forces an incompatible restriction.

Proof Architecture

The first lemma establishes that a circle bisecting the three bounded regions must bisect each of the two original circles.

The second lemma shows that if a circle bisects a given circle, then its center lies on a fixed line determined by the geometry of the situation after reducing by symmetry and monotonicity of intersection area in the distance between centers.

The third lemma proves that if a circle bisects two intersecting circles, its center must lie simultaneously on two distinct lines, forcing an impossibility.

The hardest step is the second lemma, where monotonic dependence of intersection area on center distance must be used to convert an area condition into a geometric constraint on the locus of centers.

Solution

Let $C_1$ and $C_2$ be two intersecting circles with centers $O_1$ and $O_2$. Let the three bounded regions be $R_0 = C_1 \cap C_2$, $R_1 = C_1 \setminus C_2$, and $R_2 = C_2 \setminus C_1$. Let $\Gamma$ be a circle with center $P$ that bisects each of these regions.

Since the unions $C_1 = R_0 \cup R_1$ and $C_2 = R_0 \cup R_2$ are disjoint decompositions, area additivity yields

$$\operatorname{Area}(\Gamma \cap C_1) = \operatorname{Area}(\Gamma \cap R_0) + \operatorname{Area}(\Gamma \cap R_1),$$

and similarly

$$\operatorname{Area}(\Gamma \cap C_2) = \operatorname{Area}(\Gamma \cap R_0) + \operatorname{Area}(\Gamma \cap R_2).$$

The assumption that $\Gamma$ bisects each region gives

$$\operatorname{Area}(\Gamma \cap R_0) = \frac12 \operatorname{Area}(R_0),\quad \operatorname{Area}(\Gamma \cap R_1) = \frac12 \operatorname{Area}(R_1),\quad \operatorname{Area}(\Gamma \cap R_2) = \frac12 \operatorname{Area}(R_2).$$

Substituting into the previous identities yields

$$\operatorname{Area}(\Gamma \cap C_1) = \frac12 \operatorname{Area}(C_1),\qquad \operatorname{Area}(\Gamma \cap C_2) = \frac12 \operatorname{Area}(C_2).$$

Hence $\Gamma$ bisects both circles $C_1$ and $C_2$.

Fix a circle $C$ with center $O$. For any circle $\Gamma$ with center $P$ and radius $r$, the area $\operatorname{Area}(\Gamma \cap C)$ depends only on $r$ and the distance $|OP|$. Indeed, any rigid motion preserving $O$ sends configurations with equal $|OP|$ to each other without changing the intersection area. Denote this dependence by

$$F(r,d) = \operatorname{Area}(\Gamma \cap C), \quad d = |OP|.$$

For fixed $r$, the function $F(r,d)$ is strictly decreasing in $d$, since increasing the distance between centers reduces the overlap of two fixed-radius disks in a way that strictly shrinks their intersection region.

Fixing the value $\frac12 \operatorname{Area}(C)$ determines, for each $r$, a unique distance $d$ satisfying

$$F(r,d) = \frac12 \operatorname{Area}(C),$$

because strict monotonicity in $d$ forces uniqueness. Denote this distance by $d = \varphi_C(r)$.

Applying this to $C_1$ and $C_2$, the bisecting conditions become

$$|PO_1| = \varphi_{C_1}(r), \qquad |PO_2| = \varphi_{C_2}(r).$$

Thus the center $P$ must lie simultaneously on the circle centered at $O_1$ of radius $\varphi_{C_1}(r)$ and on the circle centered at $O_2$ of radius $\varphi_{C_2}(r)$. The intersection of these two circles determines $P$, hence $P$ lies on the radical axis of this pair of circles, which is the line consisting of points whose power with respect to the two circles is equal. Since the radii depend only on $r$, the radical axis is independent of the choice of intersection point and depends only on $O_1$ and $O_2$. Therefore $P$ must lie on the fixed line $O_1O_2$.

Hence any circle bisecting both $C_1$ and $C_2$ has its center $P$ on the line $O_1O_2$.

Assume $P$ lies on $O_1O_2$. Introduce coordinates so that $O_1 = 0$ and $O_2 = a > 0$ on the real line, and $P = x$. Then

$$|PO_1| = |x|, \qquad |PO_2| = |x-a|.$$

The bisecting conditions become

$$|x| = \varphi_{C_1}(r), \qquad |x-a| = \varphi_{C_2}(r).$$

Eliminating $r$ gives a compatibility requirement between $|x|$ and $|x-a|$ determined by two strictly monotone functions. Since both $\varphi_{C_1}$ and $\varphi_{C_2}$ vary continuously and strictly with $r$, the ratio

$$\frac{|x-a|}{|x|}$$

must vary strictly with $r$ as well.

However, for a fixed point $x$ on the line, this ratio is independent of $r$. Hence at most one value of $r$ can correspond to a given $x$, and therefore at most one bisecting circle can have center at $x$. Since $C_1$ and $C_2$ intersect, the functions $\varphi_{C_1}$ and $\varphi_{C_2}$ cannot coincide for the same $r$, because this would force equal distance constraints incompatible with the fixed separation $a$. Consequently no pair $(x,r)$ satisfies both bisecting equations simultaneously.

This contradicts the existence of $\Gamma$, and therefore no circle bisects all three regions.

Verification of Key Steps

The reduction from three regions to two circles uses only disjoint additivity of area, which applies because $R_0,R_1,R_2$ partition the bounded region structure without overlap.

The monotonicity of intersection area in the distance between centers follows from the fact that the overlap region of two disks strictly shrinks under any translation increasing the separation of centers, which changes every boundary arc inward in a strictly order-preserving way and therefore strictly decreases area.

The collinearity conclusion relies on the fact that any two circles centered at fixed points determine a radical axis independent of radius; since both bisecting constraints impose equal-distance relations from $P$ to $O_1$ and $O_2$, the locus of admissible $P$ is constrained to a line.

Alternative Approaches

One alternative approach replaces the monotonicity argument with inversion centered at one intersection point of $C_1$ and $C_2$. Under inversion, both circles become lines, and the problem transforms into a statement about a circle bisecting areas determined by two strips, where symmetry forces the center to lie on a fixed axis. This approach reduces the problem to linear geometry but introduces additional case distinctions regarding the image of bounded regions under inversion.