与排序算法不同,搜索算法是比较统一的,常用的搜索除hash外仅有两种,包括不需要排序的线性搜索和需要排序的binary search。
首先介绍一下binary search,其原理很直接,不断地选取有序数组的组中值,比较组中值与目标的大小,继续搜索目标所在的一半,直到找到目标,递归算法可以很直观的表现这个描述:
int binarySearchRecursive(int A[], int low, int high, int key)
{
if (low > high) return -;
int mid = (low + high) >> ;
if (key < A[mid]) return binarySearchRecursive(A, low, mid - , key);
else if (key > A[mid]) return binarySearchRecursive(A, mid + , high, key);
else return mid;
}
但实际上,递归方法的时间效率和空间效率都不如迭代方法,迭代方法才是常用的binary search,代码如下:
int binarySearch(int A[], int low, int high, int key)
{
int mid;
while (low <= high)
{
mid = (low + high) >> ;
if (key < A[mid]) high = mid - ;
else if (key > A[mid]) low = mid + ;
else return mid;
}
return -;
}
简单计算一下Binary Search的效率:
算法流程:
1.检查上下边界--2.获取中值--3.比较--左半边进入子问题/右半边进入自问题/获得结果
1,2所需时间为常数时间,设为C。3阶段以一半的数据量重新运行函数,所以:
T(n)=T(n/2)+C
设n=2^k,则有T(2^k)=T(2^(k-1))+C=(T(2^(k-2))+C)+C=T(1)+k*C
即T(n)=log(n)*C+T(1),所以binary search是一个O(log(n))的算法。
测试函数:
void searchTest()
{
int a[] = { ,,,,,,,,,,, };
cout << binarySearch(a, , , ) << endl;
cout << binarySearch(a, , , ) << endl;
cout << binarySearch(a, , , ) << endl;
cout << endl;
cout << binarySearchRecursive(a, , , ) << endl;
cout << binarySearchRecursive(a, , , ) << endl;
cout << binarySearchRecursive(a, , , ) << endl;
}
测试结果如下:
11
-1
-1
11
-1
-1
请按任意键继续. . .
传统C函数中有bsearch这一函数,因为在现代C++中使用C库运行效率很低,加上接口并不好用,不再提及。而STL中,有以下几个关于搜索的函数。他们均作用于各个STL容器。
int count(起始迭代器,终止迭代器,key value)
return key value的数量
iterator find(起始迭代器,终止迭代器,key value)
成功:return 找到的第一个key value的迭代器
失败:return 终止迭代器
bool binary_search(起始迭代器,终止迭代器,key value)
return 是否找到
iterator lower_bound(起始迭代器,终止迭代器,key value)
return 大于或等于key value的第一个迭代器,若所有值都小于key value,返回终止迭代器
iterator upper_bound(起始迭代器,终止迭代器,key value)
return 大于key value的第一个迭代器,若所有值都小于key value,返回终止迭代器
这些函数中,count和find是作用于任意排序对象的,其效率为O(n),而binary_search, lower_bound, upper_bound是作用于有序对象的,其效率是O(logN)。
下面代码给出这些STL函数的测试:
void searchTest()
{
vector<int> b{ ,,,,,,,,,,, };
cout << "vector<int> b{ 1,2,3,4,4,7,9,11,17,20,23,39 };" << endl;
cout << "count(b.begin(), b.end(), 4):"
<< count(b.begin(), b.end(), ) << endl;
cout << endl;
cout << "find(b.begin(), b.end(), 39) - b.begin():"
<< find(b.begin(), b.end(), ) - b.begin() << endl;
cout << "find(b.begin(), b.end(), 4) - b.begin():"
<< find(b.begin(), b.end(), ) - b.begin() << endl;
cout << "find(b.begin(), b.end(), 37) - b.begin():"
<< find(b.begin(), b.end(), ) - b.begin() << endl;
cout << "find(b.begin() + 5, b.begin() + 10, 39) - b.begin():"
<< find(b.begin() + , b.begin() + , ) - b.begin() << endl;
cout << endl;
cout << "binary_search(b.begin(), b.end(), 39):"
<< binary_search(b.begin(), b.end(), ) << endl;
cout << "binary_search(b.begin(), b.end(), 37):"
<< binary_search(b.begin(), b.end(), ) << endl;
cout << endl;
cout << "lower_bound(b.begin(), b.end(), 39) - b.begin():"
<< lower_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "lower_bound(b.begin(), b.end(), 4) - b.begin():"
<< lower_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "lower_bound(b.begin(), b.end(), 37) - b.begin():"
<< lower_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "lower_bound(b.begin() + 5, b.begin() + 10, 39) - b.begin():"
<< lower_bound(b.begin() + , b.begin() + , ) - b.begin() << endl;
cout << endl;
cout << "upper_bound(b.begin(), b.end(), 39) - b.begin():"
<< upper_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "upper_bound(b.begin(), b.end(), 4) - b.begin():"
<< upper_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "upper_bound(b.begin(), b.end(), 37) - b.begin():"
<< upper_bound(b.begin(), b.end(), ) - b.begin() << endl;
cout << "upper_bound(b.begin() + 5, b.begin() + 10, 39) - b.begin():"
<< upper_bound(b.begin() + , b.begin() + , ) - b.begin() << endl;
}
测试结果:
vector<int> b{ 1,2,3,4,4,7,9,11,17,20,23,39 };
count(b.begin(), b.end(), 4):2
find(b.begin(), b.end(), 39) - b.begin():11
find(b.begin(), b.end(), 4) - b.begin():3
find(b.begin(), b.end(), 37) - b.begin():12
find(b.begin() + 5, b.begin() + 10, 39) - b.begin():10
binary_search(b.begin(), b.end(), 39):1
binary_search(b.begin(), b.end(), 37):0
lower_bound(b.begin(), b.end(), 39) - b.begin():11
lower_bound(b.begin(), b.end(), 4) - b.begin():3
lower_bound(b.begin(), b.end(), 37) - b.begin():11
lower_bound(b.begin() + 5, b.begin() + 10, 39) - b.begin():10
upper_bound(b.begin(), b.end(), 39) - b.begin():12
upper_bound(b.begin(), b.end(), 4) - b.begin():5
upper_bound(b.begin(), b.end(), 37) - b.begin():11
upper_bound(b.begin() + 5, b.begin() + 10, 39) - b.begin():10
请按任意键继续. . .