Skip to content

Instantly share code, notes, and snippets.

@jianminchen
Created January 9, 2018 06:41
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save jianminchen/3bb1990371a539a3b4ea5670df6a9da7 to your computer and use it in GitHub Desktop.
Save jianminchen/3bb1990371a539a3b4ea5670df6a9da7 to your computer and use it in GitHub Desktop.
Root of number - January 8 2018 - excellent performance
"""
1. binary search.
2. input are nonnegative. x is nonegative float and n is positive int.
3. if x in [0, 1], then search space is [0, 1]
elif x in (1, inf), then search space is (1, x)
4. if abs(mid * n - x) <= 0.001:
return mid
elif mid * n - x > 0.001:
hi = mid
else:
lo = mid
5. time complexity is O(logx)
0 - Max(1, x), 0.001, how many in total 1000 * max(1, x), log1000 Max(1, x)
not related to n, n power of number
10 * logMax(1, x)
space complexity is O(1)
if x < 1, then range should be > x
0.1 ^ 2 = 0.01
X > 1, x = 8, n = 3
root^3 = 8 => root < x
examples:
input: x = 7, n = 3
output: 1.913
input: x = 9, n = 2
output: 3
4.
"""
import math
def root(x, n):
if x == 1:
return 1
elif x < 1:
lo, hi = 0.0, 1.0
else:
lo, hi = 1.0, x * 1.0
while lo < hi:
mid = (lo + hi) / 2
diff = mid ** n - x
if abs(diff) <= 0.001:
return round(mid, 3)
elif diff > 0.001:
hi = mid
else:
lo = mid
print root(4, 2)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment