Maths Olympiad Prep

Library / /108 of 299

Geometry Difficulty 6.3 National Olympiad Prove it Iran

Let PP and PP' be two unequal regular nn-gons and AA and AA' two points inside PP and PP', respectively. Suppose {d1,d2,,dn}\{d_1, d_2, \dots, d_n\} are the distances from AA to the vertices of PP and {d1,d2,,dn}\{d'_1, d'_2, \dots, d'_n\} are the distances from AA' to the vertices of PP'. Is it possible for {d1,d2,,dn}\{d'_1, d'_2, \dots, d'_n\} to be a permutation of {d1,d2,,dn}\{d_1, d_2, \dots, d_n\}?

Solution

Suppose that P=P1P2PnP = P_1P_2 \dots P_n, P=P1P2PnP' = P'_1P'_2 \dots P'_n and that APiAP_i's are a permutation of APiA'P'_i's. Let OO and OO' be the centers of PP and PP', respectively. Without loss of generality it can be assumed that AA is inside or on the perimeter of triangle OP1P2OP_1P_2 and AP1AP2AP_1 \le AP_2; AA' is inside or on the perimeter of OP1P2O'P'_1P'_2 and AP1AP2A'P'_1 \le A'P'_2. Using the following lemma APiAP_i's and APiA'P'_i's can be sorted in ascending order.

Lemma. In a situation as described above, if nn is odd, then
AP1AP2APnAP3APn1AP4APn+321APn+32+1APn+32, \begin{aligned} & AP_1 \le AP_2 \le AP_n \le AP_3 \le AP_{n-1} \le AP_4 \le \dots \\ & \le AP_{\frac{n+3}{2}-1} \le AP_{\frac{n+3}{2}+1} \le AP_{\frac{n+3}{2}}, \end{aligned}
and if nn is even, then
AP1AP2APnAP3APn2+2APn2+1. AP_1 \le AP_2 \le AP_n \le AP_3 \le \dots \le AP_{\frac{n}{2}+2} \le AP_{\frac{n}{2}+1}.
Proof. The assertion can be easily proved by looking at the position of AA with respect to the perpendicular bisector of P2P_2 and PnP_n (which is OP1OP_1), then the perpendicular bisector of PnP_n and P3P_3, then the perpendicular bisector of P3P_3 and Pn1P_{n-1} and so on.

It can be assumed that similar inequalities hold for APiA'P'_i's. So the problem statement is equivalent to,
AP1=AP1AP2=AP2APn=APnAP3=AP3 AP_1 = A'P'_1 \le AP_2 = A'P'_2 \le AP_n = A'P'_n \le AP_3 = A'P'_3 \le \dots
Assume that PP is greater than PP'. For each 1in1 \le i \le n, (Pn+1=P1P_{n+1} = P_1, Pn+1=P1P'_{n+1} = P'_1)
APi=APi, APi+1=APi+1, PiPi+1>PiPi+1PiAPi+1>PiAPi+1. \begin{aligned} & AP_i = A'P'_i,\ AP_{i+1} = A'P'_{i+1},\ P_iP_{i+1} > P'_iP'_{i+1} \\ & \qquad \Rightarrow \angle P_iAP_{i+1} > \angle P'_iA'P'_{i+1}. \end{aligned}
Hence
360=i=1nPiAPi+1>i=1nPiAPi+1=360, 360^{\circ} = \sum_{i=1}^{n} \angle P_i AP_{i+1} > \sum_{i=1}^{n} \angle P'_i A'P'_{i+1} = 360^{\circ},
which is a contradiction. This shows that PP and PP' are equal.

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty) added by this project.