Say you have an array for which the ith element is the price of a given stock on day i.
Design an algorithm to find the maximum profit. You may complete at most k transactions.
Note:
You may not engage in multiple transactions at the same time (ie, you must sell the stock before you buy again).
A:
思路:dp
设dp[t][d]为t次交易,新进d天的最大利润:
- 每次开始的时候,初始化为不用最新一天,或者少用一次transaction.
- DP[t][d] = max(DP[t-1][d], DP[t][d-1]);
- 状态转移方程: DP[t][d] = max(DP[t][d], DP[t-1][j] + prices[d]- prices[j])
- j in [0,day-1] inclusive.
- 第一项为第i天不进行交易,第二项为枚举j 从0~i-1,第j天买入,第i天卖出
注意当k >= len(prices) / 2的时候,说明相当于无限次交易,和第122题Best Time to Buy and Sell Stock II 一样。
上面的复杂度为O(kn^2)代码如下:
class Solution { public: int maxProfit(int k, vector<int>& prices) { int n = prices.size(); if(n<=1) return 0; k = min(k, n/2); // in case given k is tooo big, vector<vector<int>> DP(k+1, vector<int>(n,0)); for(int t = 1; t < k+1; t++){// t -> transaction count for(int d =1; d < n; d ++){ // d --> day //if not use newly added value DP[t][d] = DP[t][d-1]; for(int j = d-1; j>=0; j--){ DP[t][d] = max(DP[t][d], DP[t-1][j] + prices[d]- prices[j]); } } } return DP[k][n-1]; } };然而上面的超时了。
仔细看,我们有3层循环。
然后,其实状态转移函数里。
DP[t-1][j] + prices[d]- prices[j] ==> (DP[t-1][j] - prices[j]) + price[d]
而(DP[t-1][j] - prices[j]) 每次只需要记录一个最大就可以了。
class Solution { public: int maxProfit(int k, vector<int>& prices) { int n = prices.size(); if(n<=1) return 0; k = min(k, n/2); // in case given k is tooo big, vector<vector<int>> DP(k+1, vector<int>(n,0)); for(int t = 1; t < k+1; t++){// t -> transaction count int pre_max_diff = DP[t-1][0] - prices[0]; for(int d =1; d < n; d ++){ // d --> day //if not use newly added value DP[t][d] = max(DP[t][d-1], pre_max_diff + prices[d]); pre_max_diff = max(pre_max_diff, DP[t-1][d] - prices[d] ); } } return DP[k][n-1]; } };
Mistakes:
这里不能一行一行地计算DP, 要一列一列地计算。-------这个是我以前没有这样搞过的。
也就是说。 要慎重地考虑DP哪一个 变量
No comments:
Post a Comment