斐波拉契查找及二分查找

斐波拉契查找

代码

#include <stdio.h>
#include <stdlib.h>

//求第n项的斐波拉契数
int fib(n) {
    int left = 0;
    int right = 1;
    while ( --n > 0 ) {
        right = left + right;
        left = right - left;
    } 

    return right;
}

//求n个斐波拉契数
int * fibList(n) {
    int *a = calloc(sizeof(int), n);

    for (int i = 0; i < n; i++) {
        a[i] = fib(i+1);
    }

    return a;
}

int fibSearch(int *a, int size,int search) {
    int mid, low = 0, high = size-1;
    int *p = fibList(high-low);
    while(low < high) {
        int i = 0;
        while (high - low >= p[i]) { //查找符合的斐波拉契数
            i++;
        }
        mid = low + p[i-1] -1; //构造一个中点
        printf("low=%d,high=%d\n", low, high);
        printf("p[%d]=%d,mid=%d,a[mid]=%d\n", i-1, p[i-1], mid, a[mid]);
        if ( search < a[mid] ) {
            high = mid;
        } else if (a[mid] < search) {
            low = mid+1;
        } else {
            return mid;
        }
    }

    return -1;
}

int main() {
    int a[10] = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};

    printf("%d\n", fibSearch(a, 10, 7));
}

执行过程

low=0,high=9
p[5]=8,mid=7,a[mid]=8
low=0,high=7
p[4]=5,mid=4,a[mid]=5
low=5,high=7
p[2]=2,mid=6,a[mid]=7
6

二分查找

代码

int binSearch(int *a, int size, int search) {
    int mid, low = 0, high = size-1;
    while (low < high) {
        mid = (low + high) >> 1;
        if ( search < a[mid] ) {
            high = mid;
        } else if (a[mid] < search) {
            low = mid + 1;
        } else {
            return mid;
        }
    }

    return -1;
}

说明：斐波拉契查找要优于二分查找，如上代码每次二分查找的时候向左边查找的比较次数是1，向右边查找比较次数为2。故而我们期望查找区间多落在左侧区间，少落在右侧区间，为了找到那个合适的点，最终推论出了斐波拉契的那个点（即是黄金分割点）。故而有了斐波拉契查能性能优于二分查找。

资料：https://www.bilibili.com/video/BV1db411L71m?p=56