We label 10 points on the circumference of a circle as P 1 , P 2 , … , P 1 0 . Find the number of ways to connect the points in pairs with non-intersecting chords, with no points left unconnected.
Hint:
When the number of points is 6, the number of ways is 5, as shown below.
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.
The slight difficulty with this question is to verify the pattern of Catalan numbers . Since they appear in numerous scenarios, it is useful to be able to relate them to other appearances like the restricted staircase walk not going above y = x , or the number of ways to cut a n + 2 -gon into n triangles. Finding direction bijections between these representations leads to a lot of creativity and ingenuity.
The alternative approach as laid out by Christopher is to establish the recurrence relation
C n + 1 = C 0 C n + C 1 C n − 1 + ⋯ + C n C 0 .
At times, this is much easier to justify, due to natural subcases in the counting.
When you claim that the maps are injective, you've really only proven that the mappings are functions. At least, that's what it seems to me, though any clarification would be helpful.
Log in to reply
A function is injective if no two elements in the domain map to the same element in the image.
In the first case, I stated that "This map is injective, since the up-down strokes are determined by the chords." Admittedly, I skipped a step of explanation here, since I thought this was obvious. Let me elaborate further.
If we have two distinct elements in the domain (connected points on the circumcircle), then there exists a chord in one of them, which isn't in the other. E.g. In the example I gave, one of them has a 1-4 chord while the other doesn't. Now, having an a-b chord implies the following:
1. After a, we're going up.
2. After b, we're going down.
3. The points a and b+1 are on the same level.
Especially with condition 3, we can conclude that the images (mountain ranges) are going to be distinct.
You can do the same with the second case, where I stated that "This map is injective due to the sequence of steps that we took.". Can you figure out how this works?
Suppose r n is the number of ways to connect the points in pairs with no intersecting cords, with no point left unconnected.
The main idea of my solution is to find a relation among the r 's. Pick a random starting point and connect it to the point on its immediate left. The number of ways to connect the rest is r n − 2 . Move the endpoint two points to the left. Since the chords cannot intersect each other, we now have two subproblems: number of ways to connect two points and number of ways to connect the remaining n − 4 points. This idea goes on until the endpoint coincides with the starting point.
To final answer is to sum all the values we found. Mathematically, the recurrence relation is r n = r 0 r n − 2 + r 2 r n − 4 + ⋯ + r n r 0 = ∑ i = 0 n / 2 r 2 i r n − 2 i with the base cases r 0 = 1 . What we left to do is to slowly build up r n to find the value of r 1 0 .
Great use of recursion ! I love how you simplified the problem with 10 points into problems with fewer points.
Your formula looks off - apart from that: Solution is neat :-)
what is Catalan number? '
Log in to reply
The n th Catalan number is given by C n = 2 n + 1 1 ( n 2 n ) . You can read more about it in the Catalan numbers wiki.
Yes, the answer is the 5th Catalan number. Could you explain why this is true?
If we assume that the number of ways to connect 10 points in the given way is f(10), then solution can be written as:
f(10) = 2 * f(8) + 2 * f(2) * f(6) + f(4) * f(4), where f(2) =1, f(4) = 2, and f(6)=5 and
f(8) = 2 * f(6) + 2 * f(2) * f(4) = 2x5 + 2x1x2 = 14
If n = no. of points = 2 m
then the no. of nonintersecting chords = ( m + 1 ) ( m n ) = ( 6 ) ( 5 1 0 ) = 4 2
Could you explain how you obtained that the number of non intersecting chords is m + 1 ( m n ) ?
Problem Loading...
Note Loading...
Set Loading...
By trying small cases, we recognize the pattern of 1 , 2 , 5 , 1 4 , 4 2 gives us the Catalan numbers.
We will prove that the number of ways to arrange these 2 n points is given by C n = n + 1 1 ( n 2 n ) , by setting up a bijection with the number of "mountain ranges" which are formed by n upstrokes and n downstrokes that all stay above the x − axis.
Given any set of non-intersecting chords, we will construct a mountain range in the following manner:
Starting from vertex 1, if the vertex is connected to a vertex of larger index, then we move up the mountain range. Otherwise, we will move down the mountain range. Let's see why this gives us " n upstrokes and n downstrokes that do not go below the x − axis".
The correspondence is demonstrated below:
Note: In this map, we did not use the fact that the chords are non-intersecting! In fact, there are intersecting chord diagrams that would give us the same mountain range. For example, if the vertices 1 − 3 , 2 − 4 , 5 − 6 were paired up, we will get the same mountain range as depicted in the first pair.
Now, let's find a map from the mountain ranges to the non-intersecting chords. As hinted in the previous paragraph, the potential difficulty is to ensure that the chords do not intersect.
Given any mountain range, we label the points 1 to 2 n . If the path goes up, we denote the vertex with a + . If it goes down, we denote the vertex with a − . We then connect up the lines in the following manner:
Starting from vertex 1, look for a consecutive pair of vertices that are of the form + then -. If so, we connect these up with a chord, and then ignore them. We resume looking for consecutive pairs of vertices that are of the form + then -. Let's see why this leads to n chords that do not intersect.
The correspondence is demonstrated below:
Finally, since we have 2 injective maps, we have established that the sizes of these sets are the same.
I believe that these maps form a bijection, but it's not immediately obvious to me. Any thoughts would be welcome.