How to efficiently calculate a row in pascal's triangle?
>>> def pascal(n):
... line = [1]
... for k in range(n):
... line.append(line[k] * (n-k) / (k+1))
... return line
...
>>> pascal(9)
[1, 9, 36, 84, 126, 126, 84, 36, 9, 1]
This uses the following identity:
C(n,k+1) = C(n,k) * (n-k) / (k+1)
So you can start with C(n,0) = 1
and then calculate the rest of the line using this identity, each time multiplying the previous element by (n-k) / (k+1)
.
A single row can be calculated as follows:
First compute 1. -> N choose 0
Then N/1 -> N choose 1
Then N*(N-1)/1*2 -> N choose 2
Then N*(N-1)*(N-2)/1*2*3 -> N choose 3
.....
Notice that you can compute the next value from the previous value, by just multipyling by a single number and then dividing by another number.
This can be done in a single loop. Sample python.
def comb_row(n):
r = 0
num = n
cur = 1
yield cur
while r <= n:
r += 1
cur = (cur* num)/r
yield cur
num -= 1
The most efficient approach would be:
std::vector<int> pascal_row(int n){
std::vector<int> row(n+1);
row[0] = 1; //First element is always 1
for(int i=1; i<n/2+1; i++){ //Progress up, until reaching the middle value
row[i] = row[i-1] * (n-i+1)/i;
}
for(int i=n/2+1; i<=n; i++){ //Copy the inverse of the first part
row[i] = row[n-i];
}
return row;
}