-4
$\begingroup$

Let $ABC$ be a triangle with incentre $I$. A point $P$ in the interior of the triangle satisfies $\angle PBA+\angle PCA=\angle PBC+\angle PCB$. Show that $AP\geq AI$, and the equality holds if and only if $P=I$.

Please help me. I couldn't get anything from the question..

$\endgroup$
9

2 Answers 2

1
$\begingroup$

I will only give a brief explanation to the solution of this problem.

Referring to the diagram below,

enter image description here

we need the following knowledge:- Let I be the in-center of $\triangle ABC$. The perpendicular bisector of BC and the angle bisector of $\angle A$ will meet at X and X is right on circumference of the circle ABC. In addition, X happens to be the center of the circle passing through B, I, C. Then, XB = XI = XC.

Based on the given, after some calculation, we get:-

∠IBP = ∠IBC - ∠PBC = 1/2 ∠ABC - ∠PBC = 1/2 [∠PBA - ∠PBC] …. (1)

Similarly, ∠ICP = ∠PCB - ∠ICB = ∠PCB - 1/2 ∠ACB = 1/2[∠PCB - ∠PCA] …. (2)

Since ∠PBA +∠PCA = ∠PBC +∠PCB , then ∠PBA - ∠PBC = ∠PCB - ∠PCA …. (3)

(1) , (2) and (3) imply: ∠IBP = ∠ICP and therefore BIPC is cyclic. This is equivalent to adding P as another con-cyclic point to the circle BIC. See the diagram below,

enter image description here

In $\triangle PAX$, by triangle inequality, we have $AP + PX \ge AX = AI + IX$

Result follows after subtracting $PX = IX$ from both sides.

$\endgroup$
8
  • $\begingroup$ @ Mick,. and how about $P=I$? $\endgroup$
    – pi-π
    Commented Feb 5, 2017 at 8:59
  • $\begingroup$ @NeWtoN P = I means P is right on I. From the diagram, one can see that the blue triangle is degenerated to a straight line only. Then, AP = AI. $\endgroup$
    – Mick
    Commented Feb 5, 2017 at 9:04
  • $\begingroup$ Do we need both diagrams? $\endgroup$
    – pi-π
    Commented Feb 5, 2017 at 9:06
  • $\begingroup$ @NeWtoN Two diagrams make the explanation easier to understand. $\endgroup$
    – Mick
    Commented Feb 5, 2017 at 9:08
  • $\begingroup$ @Mick: The circumcircle of ABC is a distraction, it is completely unnecessary for solving the problem. You say the circumcenter of $CBI$ "just happens" to be on the line thru $AI.$ Are you aware of some theorem that says that the angle bisector of a triangle vertex passes thru the circumcenter of the triangle defined by incenter and opposite edge? It must be fairly obscure because I couldn't find any such theorem on Wikipedia (looking in subjects such as incenter, circumcenter, angle bisector). $\endgroup$
    – J. Heller
    Commented Feb 5, 2017 at 17:09
0
$\begingroup$

The solution given at the linked website is not clear and also wrong, so I'll write a clearer and correct solution here. To really understand this, you probably need to draw a figure (as I did when writing this).

Let $J$ be the circumcenter of $\triangle CIB.$ The key part of this problem is proving that $IJ$ is parallel to $AI.$ Since $JC,$ $JB,$ and $JI$ are radii of the circumcircle, their lengths are all equal to the same value (that is, $|JC|=|JB|=|JI|=R_{CBI}$). Triangles $JCI,$ $JBI,$ and $JCB$ are isoceles, so $\angle JCI=\angle JIC,$ $\angle JBI = \angle JIB,$ and $\angle JBC=\angle JCB.$ Using the fact that $I$ is the intersection of the angle bisectors of $\triangle ABC,$ the sum of the interior angles of quadrilateral $CIBJ$ is $$ \begin{align} & 2(C/2+\angle JCB) + 2(B/2+\angle JCB) + (\pi - 2\angle JCB) \\ & \qquad = C+B + 2\angle JCB + \pi \\ & \qquad = \pi-B-A + B + 2\angle JCB + \pi \\ & \qquad = 2\pi - A + 2\angle JCB \\ & \qquad = 2\pi \end{align} $$ so $\angle JCB = A/2$ and $\angle JIB=A/2+B/2,$ which is $\pi - \angle AIB.$ So $AI$ is parallel to $IJ.$

Let $P=I.$ Then $\angle PBA=B/2,$ $\angle PCA = C/2,$ $\angle PBC=B/2,$ and $\angle PCB=C/2$ (so the condition of the problem is satisfied). When $P=I,$ $\angle PBA=\angle PBC$ and $\angle PCA=\angle PCB.$ Moving $P$ from $I$ means setting $\angle PBA = B/2-s,$ $\angle PBC=B/2+s,$ $\angle PCA=C/2-t,$ and $\angle PCB=C/2+t$ for some angle increments $s$ and $t.$ The condition $\angle PBA + \angle PCA = \angle PBC + \angle PCB$ is now equivalent to $s=-t.$ So when decreasing $\angle PBA$ by $s$ (for some $s\in (-B/2,B/2)$), $\angle PCA$ must be increased by $s.$ This means $\angle PBI = \angle PCI = s.$ So a classical geometry theorem says that quadrilateral $BPIC$ is cyclic, meaning $P$ is on the circumcircle of $\triangle CBI,$ meaning $|JP| = |JI|.$

Finally, the triangle inequality applied to $\triangle APJ$ shows that $|AP|+|PJ| \geq |AJ|=|AI|+|IJ|.$ Since $|PJ| = |IJ|,$ $|AP|\geq |AI|$ (with equality only when $\triangle APJ$ is degenerate, that is, when $P=I$).

$\endgroup$
4
  • $\begingroup$ @@J.H..Heller, Could you please provide me with a figure? $\endgroup$
    – pi-π
    Commented Feb 5, 2017 at 2:07
  • $\begingroup$ This is actually a very hard question. Based on looking at some of your more recent questions (for example, chord of length 16 in radius 10 circle), this question is a little beyond your level right now. Once you've mastered more plane geometry (circumcircles, inscribed circles, inscribed angles, cyclic quadrilaterals), you might want to come back to this question. Then it would only take a few minutes to draw your own figure on paper to follow along with the solution. $\endgroup$
    – J. Heller
    Commented Feb 5, 2017 at 2:48
  • $\begingroup$ @@J.Heller However, I know all those concepts (cyclic quadrilateral, inscribed angles, inscribed circles, etc).. But, I.just couldnot draw a fig. so, If you please provide me a figure, it would be helpful. Thanks.. $\endgroup$
    – pi-π
    Commented Feb 5, 2017 at 3:11
  • $\begingroup$ There is a difference between "knowing" and "mastering". Someone with a very good grasp of all these concepts could easily draw a figure of this problem on paper in a few minutes. It would take me longer than this to create a good quality electronic figure. $\endgroup$
    – J. Heller
    Commented Feb 5, 2017 at 3:23

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .