Dynamic Programming - 动态规划

Definition

动态规划是一种常用的优化问题求解策略。动态规划算法是一种递归算法，他的本质是通过子问题执行顺序的调度来降低算法的代价。

The core implementation of DP

Those who cannot remember the past are condemened to repeat it.

动态规划的核心是记住已经计算过的解，动态规划的方法有两种：

1. 自顶向下的备忘录法
2. 自底向上

An example for understanding DP

1. Recursion and DP

In order to calculate the value of Fibonacci, both recursion and DP are optional.

Recursion:

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

Too many repetitive computation have been done and lead to inefficiency. Fortunately, DP can avoid the redundancy computation. For each $fib(i)$, it needs to be computated once.

**However, we should notice that it will increase space complexity.**

2. Top-down variation

def fib(k):
    if k==1: return 1
    if k==0: return 0
    ans = [-1 for i in range(k+1)]
    ans[1] = 1
    ans[0] = 0
    def fcci(ans, n):
        if n==1 or n==0 or ans[n]!=-1: return ans[n]
        if ans[n-1]!=-1 and ans[n-2]!=-1:
            ans[n] = ans[n-1] + ans[n-2]
        elif ans[n-1]!=-1:
            ans[n] = ans[n-1] + fcci(ans,n-2)
        elif ans[n-2]!=-1:
            ans[n] = ans[n-2] + fcci(ans,n-1)
        else:
            ans[n] = fcci(ans,n-2) + fcci(ans,n-1)
        return ans[n]
    return fcci(ans,k)

3. Bottom-up variation

def fib(k):
    if k==0: return 0
    if k==1: return 1
    ans = [-1 for i in range(k+1)]
    ans[1] = 1
    ans[0] = 0
    for i in range(2,k+1):
        ans[i] = ans[i-1] + ans[i-2]
    return ans[k]

From the code above, we can find that the process of DP also uses the recursion.