动态规划（一）

Coin Change

转移方程：
$f(x)$：amount $x$需要的最少硬币数。

$$f(x)=min(f(x-1),f(x-2),f(x-5))+1$$
初值：

$$f(0)=0$$

def coinChange(coins, amount):
        # write your code here
        num=[float('inf')]*(amount+1)
        num[0]=0
        for i in range(amount+1):
            for money in coins:
                if i>=money:
                    num[i]=min(num[i],num[i-money]+1)
        if num[-1]==float('inf'):
            return -1

        return num[-1]

Paint House

$$\begin{align*} f_{red}(n)&=min(f_{blue}(n-1)+f_{green}(n-1))+cost_{red}\\ f_{green}(n)&=min(f_{blue}(n-1)+f_{red}(n-1))+cost_{green}\\ f_{blue}(n)&=min(f_{red}(n-1)+f_{green}(n-1))+cost_{blue} \end{align*}$$

def minCost(costs):
        # write your code here
        if not costs:
            return 0
        n=len(costs)
        minCost0=[0]*n
        minCost1=[0]*n
        minCost2=[0]*n
        for i in range(n):
            if i==0:
                minCost0[i]=costs[0][0]
                minCost1[i]=costs[0][1]
                minCost2[i]=costs[0][2]
            else:
                minCost0[i]=min(minCost1[i-1],minCost2[i-1])+costs[i][0]
                minCost1[i]=min(minCost0[i-1],minCost2[i-1])+costs[i][1]
                minCost2[i]=min(minCost0[i-1],minCost1[i-1])+costs[i][2]
        return min(minCost0[-1],minCost1[-1],minCost2[-1])

Longest Increasing Continuous Subsequence

 def longestIncreasingContinuousSubsequence(lst):
        if not lst:
            return 0
        ret_increase=[1]*len(lst) #以自己结尾的最长连续子序列
        ret_decrease=[1]*len(lst)
        for i in range(len(lst)):
            if i>=1:
               if lst[i]>lst[i-1]:
                    ret_increase[i]=ret_increase[i-1]+1
               elif lst[i]<lst[i-1]:
                   ret_decrease[i]=ret_decrease[i-1]+1
        return max(max(ret_decrease),max(ret_increase))

Counting Bits

$$f(x)=f(x>>1)+mod(x,2)$$

def countingBits(num):
    ret=[0]*(num+1)
    for i in range(1,num+1):
        ret[i]=ret[i>>2]+i%2
    return ret

Paint House 2

第$i$个房子涂第$k$种颜色的最小费用：

$$\begin{align*} &f(i,k)=min_{m\neq k}(f(i-1,m))+cost[i][k]\\ &f(0,k)=cost[0][k] \end{align*}$$

class Solution:
    """
    @param costs: n x k cost matrix
    @return: an integer, the minimum cost to paint all houses
    """
    def minCostII(self, costs):
        # write your code here

        if not costs:
            return 0
        if not costs[0]:
            return 0

        N=len(costs)
        K=len(costs[0])
        money=[[0]*(K) for _ in range(N+1)]
        for n in range(1,N+1):
            for k in range(K):
                money[n][k]=float('inf')
                for j in range(K):
                    if j!=k:
                        money[n][k]=min(money[n-1][j]+costs[n-1][k],money[n][k])

        return min(money[-1])

直接遍历的话，时间复杂度为$O(nk^2)$，运行超时。
优化算法：
计算第$i$个房子涂色需要的最小费用时，我们需要记录$f(i-1)$的最小两个值，时间复杂度变为$O(nk)$。

class Solution:
    """
    @param costs: n x k cost matrix
    @return: an integer, the minimum cost to paint all houses
    """
    def minCostII(self, costs):
        # write your code here

        if not costs:
            return 0
        if not costs[0]:
            return 0

        N=len(costs)
        K=len(costs[0])
        money=[[0]*(K) for _ in range(N+1)]

        for n in range(1,N+1):
            #找到上一轮最小的两个值
            min1=float('inf')
            min2=float('inf')
            j1=-1; j2=-1
            for k in range(K):
                if money[n-1][k]<min1:
                    min2=min1
                    j2=j1
                    min1=money[n-1][k]
                    j1=k
                elif money[n-1][k]<min2:
                    min2=money[n-1][k]
                    j2=k

            for k in range(K):
                if k!=j1:#如果k并不是上一轮最小的那个颜色，直接取最小值
                    money[n][k]=min1+costs[n-1][k]
                else:#如果k是上一轮最小的颜色，取第二小
                    money[n][k]=min2+costs[n-1][k]

        return min(money[-1])

House Robber 1

对于第$i$栋房子，有两种选择，如果不偷，那么总共偷到的最多的钱和$f(i-1)$一样，如果偷，那么总共偷到的最多的钱为$f(i-2)+A[i]$：

$$\begin{align*} f(i)=max(f(i-1),f(i-2)+A[i]) \end{align*}$$

class Solution:
    """
    @param A: An array of non-negative integers
    @return: The maximum amount of money you can rob tonight
    """
    def houseRobber(self, A):
        # write your code here
        if not A:
            return 0

        n=len(A)

        money=[e for e in A]
        for i in range(n):
            if i>=1:
                money[i]=max(money[i-1],money[i])#不偷
            if i>=2:
                money[i]=max(money[i-2]+A[i],money[i])#偷

        return money[-1]

House Robber 2

第0个和$n-1$不能同时偷。

---

参考资料：

侯卫东老师动态规划专题课程