Catalan numbers are a sequence of natural numbers that follow the formula showing below.
The first few Catalan numbers for n = 0, 1, 2, 3, … are 1, 1, 2, 5, 14, 42, 132, 429, 1430, 4862, …
Catalan numbers occur in many counting problems in combinatorics, such as count the number of ways to generate n pairs of valid parentheses, count the number of full binary trees with n+1 leaves. Given a number n , you can find n-th Catalan number using the following 3 implementations.
Table of Content
Iteration
This solution is applying the following alternative expression of Cn. It can be implemented with iteration. It has the best time complexity O(n).
Java
|
1 2 3 4 5 6 7 8 9 |
//Iteration, using formula C0=1 Cn+1= 2(2n+1)/(n+2)*Cn //Time O(n) Space O(1), n is input number public static long catalanIte(int n) { long res = 1; for (int i = 0; i < n; i++) { res = res * 2*(2*i+1) / (i+2); } return res; } |
Javascript
|
1 2 3 4 5 6 7 8 9 |
//Iteration, using formula C0=1 Cn+1= 2(2n+1)/(n+2)*Cn //Time O(n) Space O(1), n is input number function catalanIte(n) { var res = 1; for (let i = 0; i < n; i++) { res = res *2*(2*i+1) / (i+2); } return res; } |
Python
|
1 2 3 4 5 6 7 |
# Iteration, using formula C0=1 Cn+1= 2(2n+1)/(n+2)*Cn # Time O(n) Space O(1), n is input number def catalanIte(n) : res = 1 for i in range(0, n) : res = res *2*(2*i+1) / (i+2) return int(res) |
Recursion
This solution is using the another alternative expression of Cn. It can be implemented with recursion. It has the worst time complexity O(2^n) with repeated calls.
Java
|
1 2 3 4 5 6 7 8 9 10 11 |
//Recursion, formula C0=1, Cn = C0*Cn-1+C1*Cn-2..+ Cn-i*C0 //Time O(2^n), Space O(2^n) public static long catalanRec(int n) { if (n <= 1) // Base condition return 1; int res = 0; for (int i = 0; i < n; i++) { res += catalanRec(i) * catalanRec(n-i-1); } return res; } |
Javascript
|
1 2 3 4 5 6 7 8 9 10 11 |
//Recursion, formula C0=1, Cn = C0*Cn-1+C1*Cn-2..+ Cn-i*C0 //Time O(2^n), Space O(2^n) function catalanRec(n) { if (n <= 1) // Base condition return 1; var res = 0; for (let i = 0; i < n; i++) { res += catalanRec(i) * catalanRec(n-i-1); } return res; } |
Python
|
1 2 3 4 5 6 7 8 9 |
# Recursion, formula C0=1, Cn = C0*Cn-1+C1*Cn-2..+ Cn-i*C0 # Time O(2^n), Space O(2^n) def catalanRec(n): if n <= 1: return 1 res = 0 for i in range(n): res += catalanRec(i) * catalanRec(n-i-1) return res |
Doodle
Dynamic programming
This solution uses the same expression as recursion. But it overcomes the overlapping in recursion by using dynamic programming technique tabulation. The result from previous calls are saved to a 2d array. They can be re-used for the following steps without calling the recursion. It improves the time complexity from exponential to O(n^2).
Java
|
1 2 3 4 5 6 7 8 9 10 11 12 13 |
//Dynamic programming, Time O(n^2), Space O(n) public static long catalanDP(int n) { int memo[] = new int[n+1]; memo[0] = 1; memo[1] = 1; for (int i = 2; i <= n; i++) { memo[i] = 0; for (int j = 0; j < i; j++) { memo[i] += memo[j] * memo[i-j-1]; } } return memo[n]; } |
Javascript
|
1 2 3 4 5 6 7 8 9 10 11 12 13 |
//Dynamic programming, Time O(n^2), Space O(n) function catalanDP(n) { var memo = []; memo[0] = 1; memo[1] = 1; for (let i = 2; i <= n; i++) { memo[i] = 0; for (let j = 0; j < i; j++) { memo[i] += memo[j] * memo[i-j-1]; } } return memo[n]; } |
Python
|
1 2 3 4 5 6 7 8 9 10 |
#Dynamic programming, Time O(n^2), Space O(n) def catalanDP(n) : memo = [0]*(n+1) memo[0] = 1 memo[1] = 1 for i in range(2, n+1) : memo[i] = 0 for j in range(0, i): memo[i] += memo[j] * memo[i-j-1] return memo[n] |
Download Java, JavaScript and Python code
Catalan numbers animated visual
Coding interview questions of parentheses and Catalan number