斐波那契数列——普通递归、记忆化搜索、动态规划

普通递归

import time


def fibo(n):
    if n == 0:
        return 0
    if n == 1:
        return 1
    return fibo(n-1) + fibo(n-2)


n = 35

start_time = time.time()
a = fibo(n)
end_time = time.time()

print("fibo(%d) = %d" % (n, a))
print("runing time = ", end_time-start_time)

普通递归由于进行了大量重复计算，n=35在我的计算机上需要5秒多，n每增加1运行时间为指数级增长。

记忆化搜索

import time


def fibo(n):
    if n == 0:
        return 0
    if n == 1:
        return 1
    if memory[n] == -1:
        memory[n] = fibo(n-1) + fibo(n-2)
    return memory[n]


n = 35
memory = [-1] * (n+1)

start_time = time.time()
a = fibo(n)
end_time = time.time()

print("fibo(%d) = %d" % (n, a))
print("runing time = ", end_time-start_time)

记忆化搜索，建立一个列表存储计算过的数值，递归的时候不需要重复计算，运行时间大大减少。由于python递归存在最大递归深度，当n数值较大时，例如n=10000，会抛出异常：RecursionError: maximum recursion depth exceeded in comparison.

动态规划

import time


def fibo(n):
	memory = [0, 1]
	for i in range(2, n+1):
		memory.append(memory[i-1] + memory[i-2])
	return memory[n]


n = 10000

start_time = time.time()
a = fibo(n)
end_time = time.time()

print("fibo(%d) = %d" % (n, a))
print("runing time = ", end_time-start_time)