算法之二分查找

概念

二分查找针对的是一个有序的数据集合，查找思想有点类似分治思想，每次都通过跟区间的中间元素对比，将待查找的区间缩小为之前的一半，直到找到想要查找的元素，或是区间被缩小为0。

时间复杂度

二分查找的时间复杂度为$O(logn)$

$logn$是一个非常恐怖的数量级，即使n非常大，对应的$logn$也很小。比如n等于2的32次方，这个数很大，大约42亿。也就是说，我们在42亿个数据中用二分查找进行查找，查找到一个元素只需要比较32次。

二分查找的递归实现

Python：

def main():
        s_raw = input()
        s_string = s_raw.split(",")
        s_num = [int(x) for x in s_string]
        v = int(input())
        print bsearch(s_num, 0, len(s_num), v)

def bsearch(s, low, high, v):
        mid = (low + high)/2
        if s[mid] == v:
                return mid
        elif v < s[mid]:
                high = mid
                return bsearch(s, low , high, v)
        else:
                low = mid
                return bsearch(s, low, high, v)

main()

二分查找应用场景

1. 静态：数据插入和删除操作不频繁
2. 连续：数据必须是排序好的
3. 数组存储：必须使用数组存储而非链表存储
4. 大小刚好合适：太大的情况下可能在内存中难以找到一块连续的地址空间，太小的话还不如使用遍历方便。

二分查找的几个变形问题

1.查找第一个值等于给定值的元素

def main():
        s_raw = input()
        s_list = s_raw.split(",")
        s = [int(x) for x in s_list]

        v = int(input())
        low = 0
        high = len(s) - 1
        print(bsearch(s, low, high, v))

def bsearch(s, low, high, v):
        while(low <= high):
                mid = (low + high) / 2
                if s[mid] > v:
                        high = mid - 1
                elif s[mid] < v:
                        low = mid + 1
                else:
                        if mid == 0 or s[mid - 1] != v:
                                return mid
                        else:
                                high = mid - 1
        return -1

main()

2.查找最后一个值等于给定值的元素

def main():
        s_raw = input()
        s_list = s_raw.split(",")
        s = [int(x) for x in s_list]

        v = int(input())
        low = 0
        high = len(s) - 1
        print(bsearch(s, low, high, v))

def bsearch(s, low, high, v):
        while(low <= high):
                mid = (low + high) / 2
                if s[mid] > v:
                        high = mid - 1
                elif s[mid] < v:
                        low = mid + 1
                else:
                        if mid == len(s)-1 or s[mid + 1] != v:
                                return mid
                        else:
                                low = mid + 1
        return -1

main()

3.查找第一个大于等于给定值的元素

def main():
        s_raw = input()
        s_list = s_raw.split(",")
        s = [int(x) for x in s_list]

        v = int(input())
        low = 0
        high = len(s) - 1
        print(bsearch(s, low, high, v))

def bsearch(s, low, high, v):
        while(low <= high):
                mid = (low + high) / 2
                if s[mid] < v:
                        low = mid + 1
                else:
                        if mid == 0 or s[mid - 1] < v:
                                return mid
                        else:
                                high = mid - 1
        return -1

main()

4.求循环有序数组如4、5、6、1、2、3中的某个数（二分查找）

def main():
        s_raw = input()
        s_list = s_raw.split(",")
        s = [int(x) for x in s_list]

        v = int(input())
        low = 0
        high = len(s) - 1
        print(bsearch(s, low, high, v))


def bsearch(s, low, high, v):
        while(low <= high):
                mid = (low + high) / 2
                if s[mid] > v:
                        if s[low] > v:
                                low = mid + 1
                        else:
                                high = mid - 1
                elif s[mid] < v:
                        low = mid + 1
                else:
                        return mid
main()