Cutting regular polygon

Geometry Level 5

Find the sum of all positive integers n 3 n\ge 3 such that it is possible to cut a regular n n -gon into several smaller pieces, each of which is a (not necessarily congruent) regular polygon .


The answer is 25.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

Mark Hennings
Dec 9, 2016

Obviously, the triangle (into four triangles) and the square (into four squares) can be dissected as required. Let us restrict our attention to n n -gons with n > 4 n > 4 , so that the interior angles are obtuse and the exterior angles acute.

If, in a dissection, one of the angles of the n n -gon is not "cut", then an n n -gon must be used to fill that angle. Since more than one subpolygon is required, this means that the two sides adjacent to that corner must be "cut", and hence that the exterior angle of the n n -gon must be fillable by the interior angles of a set of regular polygons. Since this exterior angle is acute (and the only acute interior angle of a regular polygon is 6 0 60^\circ ),. we deduce that the exterior angle of the polygon must be 6 0 60^\circ , and hence n = 6 n=6 , so that we have the hexagon, which can be dissected into hexagons and triangles each of a third the original side length, as shown on the right.

Thus, otherwise, every angle of the regular n n -gon must be "cut", and so the (obtuse) interior angle of the regular n n -gon must be splittable into a sum of two or more interior angles of regular polygons. There are only three options for expressing an obtuse angle as the sum of two or more regular polygon interior angles, namely 12 0 = 6 0 + 6 0 120^\circ = 60^\circ + 60^\circ , 15 0 = 6 0 + 9 0 150^\circ = 60^\circ + 90^\circ and 16 8 = 10 8 + 6 0 168^\circ = 108^\circ + 60^\circ .

  • The first of these gives the hexagon, dissected into six triangles.
  • The second of these gives the dodecagon, dissected into six squares, six triangles and one hexagon.
  • The third forces us to consider the regular 30 30 -gon, with each angle split into a pentagon and a triangle. The exterior angle of a pentagon cannot be split into a sum of interior angles of regular polygons, and so the pentagon cannot "cut" any side of the n n -gon it touches. Thus the side-length of each pentagon must be the side-length of the original n n -gon. The only way this is possible is for alternate faces of the 30 30 -gon to be sides of pentagons, and for the other faces to be sides of triangles. These triangles could be further subdissected, but that is not important. But, "one layer in", this creates angles of 36 0 6 0 2 × 10 8 = 8 4 360^\circ - 60^\circ - 2\times108^\circ = 84^\circ , which cannot be filled by the interior angles of regular polygons. This option is not possible.

Thus the answer is 3 + 4 + 6 + 12 = 25 3+4+6+12 = \boxed{25} .

Oh wow, this classification makes the problem so much easier! Smart way of dealing with our concerns about the "cut vertex" (see discussion on Geoff's solution).

I added the image of "hexagons and triangle each a third of the orginal side length", taken from Geoff's inspired problem .

Calvin Lin Staff - 4 years, 6 months ago

Log in to reply

Appropriate graphic. Shame about the colour scheme!

Mark Hennings - 4 years, 6 months ago

Log in to reply

What about the color scheme?

I can update it if desired.

Calvin Lin Staff - 4 years, 6 months ago

Log in to reply

@Calvin Lin I can live with it. But that blue and that fuchsia! These colours are part of Windows history, when there were only a seriously limited number of colour options available! In this day and age, there are infinitely more aesthetically pleasing options.

Mark Hennings - 4 years, 6 months ago

Log in to reply

@Mark Hennings Alright, I waved my magic wand and the colors changed :)

Calvin Lin Staff - 4 years, 6 months ago

Log in to reply

@Calvin Lin Much more pleasing to the eye, thanks!

Mark Hennings - 4 years, 6 months ago

It's kinda disappointing that dissecting a 30-gon is impossible. It would have made a nice non-trivial solution.

Julian Poon - 4 years, 6 months ago

I almost got it, but I didn't think about why an octagon doesn't work.

Guy Alves - 4 years, 6 months ago
Geoff Pilling
Dec 8, 2016

I love this problem... Thanks for posting Khang!

Here is my rough solution:

The following regular polygons can be further divided into regular polygons:

  • Triangle: four triangles
  • Square: four squares
  • Hexagon: six triangles
  • 12-gon: six triangles, six squares, and a hexagon

Here is my first cut argument as to why none of the remaining polygons don't work...

Cuts have to go through either the vertices or the sides or both. I persuaded myself that only the triangle and square can have cuts going through the sides. (Will try to come up with a more concrete proof)

For the remainder of the solids (cuts going through the vertices) the angle would have to add up to angles of other regular polygons.

The interior angles of the first few regular polygons are:

  • Triangle: 60 degrees
  • Square: 90 degrees
  • Pentagon: 108 degrees
  • Hexagon: 120 degrees

(No need to go beyond the hexagon since the limit on the interior angle of any n-gon is 180 degrees, so the remaining angle after a cut would be less than 60 degrees, the angle of the smallest n-gon)

So, only regular polygons whose interior angles add up to these numbers can work. That includes:

  • Hexagon: 120 degrees (60 + 60)
  • 12-gon: 150 degrees (90 + 60)

That leaves the four regular polygons I mentioned at the beginning.

So, the answer is 3 + 4 + 6 + 12 = 25 3+4+6+12=\boxed{25}

Make sense?

That argument doesn't rule out a 30-gon (internal angle of 168 degrees, can be split into 108 + 60).

Ivan Koswara - 4 years, 6 months ago

I also started with the "vertex angle" consideration of a valid construction. However, you miss the case of using a smaller version of the n-gon, which allows us to fill in the angle completely. The square cut into 4 squares is an excellent example that this case could happen and still lead to a valid result.

Of course, this delays the angle consideration to a later point, but then we have lesser control over what happens there. Issues that I ran into was that the 2 side angles could be combined at a point.

Calvin Lin Staff - 4 years, 6 months ago

Log in to reply

Yeah, the only polygons I could think of where the cuts are through sides are the triangle and the square.

Geoff Pilling - 4 years, 6 months ago

Log in to reply

The tiling of the hexagon by equilateral triangles of side length s/2 also gives us cuts through sides.

Disclaimer: I am equally puzzled about the rigorous proof.

Calvin Lin Staff - 4 years, 6 months ago

Log in to reply

@Calvin Lin Ah, true good point! I should have clarified... "The only polygons where we rely on cutting through sides, i.e. only answers are through the sides" :)

Geoff Pilling - 4 years, 6 months ago

@Calvin Lin Yeah, I'll think about it and see if I can come up for a rigorous proof... For now I'd only want to advertise this as a reasonable argument....

Geoff Pilling - 4 years, 6 months ago

@Calvin Lin Also, a hexagon can be side-cut into a combination of hexagons and triangles of side length s / 3 s/3 .

Mark Hennings - 4 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...