斐波那契数列
实现斐波那契数列
0、1、1、2、3、5、8、13、21、34、……
F(0)=0,F(1)=1,F(n)=F(n-1)+F(n-2)(n≥2,n∈N*)
递归实现:
package org.fan.learn.dp;
/**
* Created by fan on 2016/9/12.
*/
public class Fibo {
//注意这里是long型的,斐波那契数列很大
public static long[] result;
public static final int MAX = 100000;
public static long fibo(int idx) {
if (idx <= 0) {
return 0;
}
if (idx == 1) {
return 1;
}
//记忆化搜索
if (result[idx] >= 0) {
return result[idx];
}
result[idx] = fibo(idx-1) + fibo(idx-2);
return result[idx];
}
public static void main(String[] args) {
long start = System.currentTimeMillis();
result = new long[MAX];
for (int i = 0; i < MAX; i++) {
result[i] = -1;
}
System.out.println(fibo(50));
System.out.println("Time:" + (System.currentTimeMillis() - start) + " ms");
/*
不记忆化搜索:
12586269025
Time:124053 ms
记忆化搜索:
12586269025
Time:3 ms
*/
}
}
如果不加记忆化搜索,时间复杂度为:n!
加了记忆化搜索,时间复杂度为:n
不加记忆化搜索的时间复杂度怎么计算的?
比如计算f(idx)需要计算f(idx-1),f(idx-2),f(idx-3)…….f(2),f(1)
而计算f(idx-1)时,又要重复计算f(idx-2),f(idx-3)…….f(2),f(1)
计算f(idx-2)时,又要重复计算f(idx-3)…….f(2),f(1)
这样f(1)重复计算了n次,f(2)重复计算了n-1次(因为计算f(1)时无需计算f(2)),以此类推,时间复杂度为O(n!)。
加了记忆化搜索之后,没有了重复计算,因此是O(n)
递推实现:
//递推实现
public static long fibo2(int idx) {
result[0] = 0;
result[1] = 1;
for (int i = 2; i <= idx; i++) {
result[i] = result[i-1] + result[i-2];
}
return result[idx];
}
完整代码:
package org.fan.learn.dp;
/**
* Created by fan on 2016/9/12.
*/
public class Fibo {
public static long[] result;
public static final int MAX = 100000;
public static long fibo(int idx) {
if (idx <= 0) {
return 0;
}
if (idx == 1) {
return 1;
}
//记忆化搜索
if (result[idx] >= 0) {
return result[idx];
}
result[idx] = fibo(idx-1) + fibo(idx-2);
return result[idx];
}
//递推实现
public static long fibo2(int idx) {
result[0] = 0;
result[1] = 1;
for (int i = 2; i <= idx; i++) {
result[i] = result[i-1] + result[i-2];
}
return result[idx];
}
public static void main(String[] args) {
long start = System.currentTimeMillis();
result = new long[MAX];
for (int i = 0; i < MAX; i++) {
result[i] = -1;
}
System.out.println(fibo(50));
System.out.println("Digui Time:" + (System.currentTimeMillis() - start) + " ms");
/*
不记忆化搜索:
12586269025
Time:124053 ms
记忆化搜索:
12586269025
Time:3 ms
*/
long start2 = System.currentTimeMillis();
for (int i = 0; i < MAX; i++) {
result[i] = -1;
}
System.out.println(fibo2(50));
System.out.println("Ditui Time:" + (System.currentTimeMillis() - start2) + " ms");
}
}
执行结果:
12586269025
Digui Time:10 ms
12586269025
Ditui Time:1 ms
可以肯定的是,递推要比递归的时间用的要少,即使递归有记忆化搜索的优化。
一个参考的菲波那切数列数列在线计算器,验证代码的正确性:
http://www.99cankao.com/algebra/fibonacci.php
n!
5!=5*4*3*2*1=120
代码实现:
package org.fan.learn.dp;
/**
* Created by fan on 2016/9/12.
*/
public class Multiply {
public static long fDigui(int idx) {
if (idx == 1) {
return 1;
}
return fDigui(idx-1) * idx;
}
public static long fDitui(int idx) {
if (idx < 1) {
return 0;
}
long result = 1L;
for (int i = 1; i <= idx; i++) {
result = result * i;
}
return result;
}
public static void main(String[] args) {
long start = System.currentTimeMillis();
System.out.println(fDigui(10));
System.out.println("Digui Time:" + (System.currentTimeMillis() - start) + " ms");
long start2 = System.currentTimeMillis();
System.out.println(fDitui(10));
System.out.println("Ditui Time:" + (System.currentTimeMillis() - start2) + " ms");
}
}
注意:
n!这个的递归实现是没有冗余的。
运行结果:
3628800
Digui Time:1 ms
3628800
Ditui Time:0 ms