斐波那契查找算法的实现（又称黄金分割查找法；【java语言实现】）

核心思想

由斐波那契数列 F[k]=F[k-1]+F[k-2] 的性质，可以得到 （F[k]-1）=（F[k-1]-1）+（F[k-2]-1）+1 。
该式说明：只要顺序表的长度为F[k]-1，则可以将该表分成长度为F[k-1]-1和F[k-2]-1的两段，从而中间位置为mid=low+F(k-1)-1
将mid作为顺序表的分割点
由于mid的确认过程不涉及除法，仅仅需要进行加或者减；理论上比二分查找法消耗要小，速度快

实现代码

package com.zhaojinlei.test;

import java.util.Arrays;

/**
 * className:FibonacciSearch
 *
 * @author:ZHAOJINLEI
 * @version:0.1
 * @date:2020/9/517:17
 * @since:jdk1.8
 */
public class FibonacciSearch {

    public static void main(String[] args) {
        //测试1
        int[] ser = {0,3,5,9,12,13,26,66,99};
        int index = search(ser, 26);
        System.out.println("{0,3,5,9,12,13,26,66,99}中查找26：");
        System.out.println("index:"+index +"  value:"+ser[index]);
        System.out.println();
        //测试2
        int[] ser1 = {0,3,5,9,12,13,26,66,99,109,110,112,116,225,336,995,1000,1001,1002,1030};
        int index1 = search(ser1, 995);
        System.out.println("{0,3,5,9,12,13,26,66,99,109,110,112,116,225,336,995,1000,1001,1002,1030}中查找995：");
        System.out.println("index:"+index1 +"  value:"+ser1[index1]);

    }

    //创建一个含n个数的斐波那契数列
    public static int[] createFibonacci(int n) {
        int[] ints = new int[n];
        ints[0] = 1;
        ints[1] = 1;

        for (int i = 2; i < ints.length; i++) {
            ints[i] = ints[i - 1] + ints[i - 2];
        }
        return ints;
    }

    public static int search(int ser[], int elem) {
        int n = 10;  //初始斐波那契数列长度10
        int[] fibs = createFibonacci(n);
        while (fibs[fibs.length - 1] < ser.length) {//斐波那契数列长度不够，进行扩容（每次扩2倍）
            n *= 2;
            fibs = createFibonacci(n);
        }
        int low = 0;
        int high = ser.length;
        int k = fibs.length - 1;
        int mid;


        while (fibs[k] - 1 > ser.length) {//找到符合要求的k的最小值
            k--;
        }
        k++;
        int[] ints = Arrays.copyOf(ser,fibs[k]-1);//复制原数组

        for (int i = ser.length; i < ints.length; i++) {//超出原数组的最大索引的位置填充原数组最大值
            ints[i] = ser[ser.length - 1];
        }
        while (low < high) {

            //检查两端
            if(ints[low] == elem)
                return low;
            if(ints[high] == elem)
                return high;

            //黄金分割点
            mid = low + fibs[k - 1];

            if (ints[mid] == elem){
                if(mid<ser.length-1){
                    return ser.length-1;
                }else {
                    return mid;
                }
            }
            //低区查找
            if(ints[mid]>elem){
                high = mid-1;
                k--;
            }
            //高区查找
            if(ints[mid]<elem){
                low = mid+1;
                k-=2;
            }
        }

        return -1;
    }

}

测试结果