lintcode:Find Peak Element_find peak element lintcode-CSDN博客

本文链接：https://blog.csdn.net/gao1440156051/article/details/49907215

There is an integer array which has the following features:

The numbers in adjacent positions are different.
A[0] < A[1] && A[A.length - 2] > A[A.length - 1].

We define a position P is a peek if:

A[P] > A[P-1] && A[P] > A[P+1]

Find a peak element in this array. Return the index of the peak.

Example
Given [1, 2, 1, 3, 4, 5, 7, 6]

Return index 1 (which is number 2) or 6 (which is number 7)

Note
The array may contains multiple peeks, find any of them.

Challenge
Time complexity O(logN)

由时间复杂度的暗示可知应使用二分搜索。首先分析若使用传统的二分搜索，若A[mid] > A[mid - 1] && A[mid] < A[mid + 1]，则找到一个peak为A[mid]；若A[mid - 1] > A[mid]，则A[mid]左侧必定存在一个peak，可用反证法证明：若左侧不存在peak，则A[mid]左侧元素必满足A[0] > A[1] > … > A[mid -1] > A[mid]，与已知A[0] < A[1]矛盾，证毕。同理可得若A[mid + 1] > A[mid]，则A[mid]右侧必定存在一个peak。如此迭代即可得解。

备注：如果本题是找 first/last peak，就不能用二分法了

class Solution {
public:
    /**
     * @param A: An integers array.
     * @return: return any of peek positions.
     */
    int findPeak(vector<int> A) {
        // write your code here
        int l=0,r=A.size()-1;
        //l,r-1,r是循环的最小情况
        while(l<r-1){
            int mid=l+(r-l)/2;
            if(A[mid]>A[mid-1]&&A[mid]>A[mid+1]){
                return mid;
            }else if(A[mid]<A[mid-1]){
                r=mid;
            }else if(A[mid]<A[mid+1]){
                l=mid;
            }
        }
    }
};