Tony's Log

Algorithms, Distributed System, Machine Learning

  博客园 :: 首页 :: 博问 :: 闪存 :: 新随笔 :: 联系 :: 订阅 订阅 :: 管理 ::

One sentence in the problem statement is crucial "your friend will also play optimally", if you interpret it correctly, you are very close to AC. What does it mean? It means, no matter what your current choice is, 1 brick or 2 or 3, your opponent will take optimized score of the rest bricks. Then, it is natural to have a 'reverse' DP:

(thanks to http://massivealgorithms.blogspot.com/2015/01/hackerrank-play-game.html)

t = int(input())
for i in range(t):
    n = int(input())
    arr = [int(i) for i in input().strip().split()]
    
    ######
    def calc(arr):
        alen = len(arr)
        if alen < 4:
            return sum(arr)
        '''
        Both DP[] and PreSum[] is in terms of bottom-up
        because opponent 'your friend will also play optimally'
        '''
        #    Calc prefix sum
        presum = [0] * alen
        presum[0] = arr[0]
        for i in range(1, alen):
            presum[i] = presum[i - 1] + arr[i]
            
        #    Go DP
        dp = [0] * (alen)        
        dp[0] = arr[0]
        dp[1] = arr[1] + dp[0]
        dp[2] = arr[2] + dp[1]
        for i in range(3, alen):
            # Take 1: (i), opponent will take dp[i - 1]
            x = arr[i] + presum[i - 1] - dp[i - 1]
            # Take 2
            y = presum[i - 2] + arr[i] + arr[i - 1] - dp[i - 2]
            # Take 3
            z = presum[i - 3] + arr[i] + arr[i - 1] + arr[i - 2] - dp[i - 3]
            dp[i] = max(x, y, z)
            
        return dp[alen - 1]
    ######    
    # reverse to bottom -> top
    arr.reverse()
    print (calc(arr))
posted on 2015-03-25 06:24  Tonix  阅读(686)  评论(0编辑  收藏  举报