Troika of Google big data processing: GFS, Bigtable and MapReduce
The main idea of MapReduce:
Divide the original problem into several sub-problems recursively, until the sub-problems meet the boundary conditions, stop the recursion. Break down the sub-problems one by one, combine the solved sub-problems, and finally the algorithm will merge layer by layer to get the answer to the original problem
The steps of the divide and conquer algorithm:
- Divide: Recursively decompose the problem into sub-problems
- Governance: break down these smaller sub-problems one by one
- Combine: Merge the solved sub-problems layer by layer, and finally get the solution of the original problem
Situations where divide and conquer applies
- The computational complexity of the original problem increases with the size of the problem
- The original problem can be broken down into smaller sub-problems
- The structure and nature of the sub-problems are the same as the original problem, and they are independent of each other. The sub-problems do not contain common sub-sub-problems
- The solutions of the subproblems decomposed from the original problem can be combined into the solution of the problem
algorithm
python code:
class Solution:
def majorityElement(self, nums: List[int]) -> int:
if len(nums) == 1:
return nums[0]
left = self.majorityElement(nums[:len(nums)//2])
right = self.majorityElement(nums[len(nums)//2:])
if left == right:
return right
elif nums.count(right) > nums.count(left):
return right
else:
return left
python code
class Solution:
def maxSubArray(self, nums: List[int]) -> int:
if len(nums) == 1:
return nums[0]
left = self.maxSubArray(nums[:len(nums)//2])
right = self.maxSubArray(nums[len(nums)//2:])
max_l = nums[len(nums)//2]
tmp = 0
for i in range(len(nums)//2,len(nums)):
tmp += nums[i]
max_l = max(tmp,max_l)
max_r = nums[len(nums)//2-1]
tmp_1 = 0
for i in range(len(nums)//2-1,-1,-1):
tmp_1 += nums[i]
max_r = max(tmp_1,max_r)
return max(left,right,max_l+max_r)
python code
class Solution:
def myPow(self, x: float, n: int) -> float:
if n < 0 :
x = 1/x
n = -n
if n == 0:
return 1
if n%2 == 1:
p = x*self.myPow(x, n-1)
return p
return self.myPow(x*x,n/2)
References
One of the five commonly used algorithms: divide and conquer algorithm