Pattern visualizer
Koko Eating Bananas
There's no array of speeds to search — but speed behaves monotonically: a faster k always needs fewer-or-equal hours, never more. That monotonic relationship is exactly what binary search needs, so it works the same as searching a sorted array, just with 'is this speed fast enough?' standing in for 'is this the target?'. Animated on: piles = [3,6,7,11], h = 8 — find the minimum constant eating speed k so all piles finish within h hours..
Binary search the ANSWER (speed), not the array
time O(n log(max(piles)))space O(1)step 1 / 5
3
[0]6
[1]7
[2]11
[3]line 2
piles=[3,6,7,11], h=8 hours available. Binary search the SPEED k (1..11): higher k always means fewer or equal hours needed, so hours-needed is monotonic in k — perfect for binary search.
Pseudocode
1FUNCTION minEatingSpeed(piles, h):2 low = 1, high = the largest pile3 WHILE low <= high:4 mid = (low + high) / 2, rounded down5 hours = for each pile, add up ceil(pile / mid)6 IF hours <= h:7 ans = mid; high = mid - 1 (try slower)8 ELSE:9 low = mid + 1 (too slow, go faster)10 RETURN ans
← / → step · space play · Home restart