• <ins id="pjuwb"></ins>
    <blockquote id="pjuwb"><pre id="pjuwb"></pre></blockquote>
    <noscript id="pjuwb"></noscript>
          <sup id="pjuwb"><pre id="pjuwb"></pre></sup>
            <dd id="pjuwb"></dd>
            <abbr id="pjuwb"></abbr>

            superman

            聚精會(huì)神搞建設(shè) 一心一意謀發(fā)展
            posts - 190, comments - 17, trackbacks - 0, articles - 0
               :: 首頁 :: 新隨筆 :: 聯(lián)系 :: 聚合  :: 管理

            ZOJ 1102 - Phylogenetic Trees Inherited

            Posted on 2008-04-06 11:13 superman 閱讀(430) 評(píng)論(0)  編輯 收藏 引用 所屬分類: ZOJ

            Official Solution:

            Problem G: Phylogenetic Trees Inherited

            The first thing to observe is that the different positions in every sequence are independent of each other. This reduces the tree of sequences to a tree of amino acids. At the root of the tree, or for that matter of any sub-tree, there may be many possible amino acids leading to optimal costs. Suppose, you have calculated for two sub-trees Tl and Tr the sets of amino-acids leading to optimal costs Al and Ar. Adjacent sub-trees Tl and Tr have as their father the node T. Now you want to find the set of amino-acids A that you can mark T with, leading to optimal costs for T.

            There are two cases: if the intersection of Al with Ar is non-empty, define A as just this intersection, otherwise define A to be the union of Al and Ar. To see why this is true, observe the extra costs you get, when you assemble T from Tl and Tr. In the first case, you have 0 extra costs when you take an amino-acid from the intersection, but 1 or 2 extra costs when you do not. In the second case, you have 1 extra costs when you take an amino-acid from the union, but 2 extra-costs when you do not. Now, you may want to assemble T not from Tl and Tr but from some other sub-optimal trees. As you can easily verify, this leads to sub-optimal costs for T as well.

            This reasoning is carried over straightforwardly to an induction proof and leads to a dynamic programming solution. Since the amino-acids are upper-case letters, you can represent sets of amino-acids as ints. The set operations you need are then easily performed as bitwise and respectively or. Whenever you do a union operation, your costs increase by 1.

            Another, more straight-forward solution is to calculate for each node of the tree the optimal costs for every amino acid the node can be marked with. This is done by trying every possible combination of amino acids for the two sub-trees, assuming their optimal costs have already been calculated. Since this solution might turn out to be too inefficient, it can be improved upon by observing that a father node always can be marked with either the left or the right son's amino-acid - there is no need to take an amino acid that differs from both.

            Judges' test data was constructed from a test-case with a few long sequences, a test-case with many short sequences, a test-case where every possible pair of amino-acids occured, and 100 random-generated test-cases where length and number of sequences are geometrically distributed. The total number of test-cases is 110. Since there may be multiple correct answers for the test cases, a special verification program was written by slightly modifying the judges' solution.


             1 /* Accepted 1102 C++ 00:00.56 1040K */
             2 #include <string>
             3 #include <iostream>
             4 
             5 using namespace std;
             6 
             7 int main()
             8 {
             9     int n, l;
            10     while((cin >> n >> l) && n && l)
            11     {
            12         int heap[2048], cost = 0;
            13         string seq[1024];
            14         
            15         for(int i = 0; i < n; i++)
            16             cin >> seq[i];
            17         
            18         for(int i = 0; i < l; i++)
            19         {
            20             for(int k = 0; k < n; k++)
            21                 heap[n + k] = 1 << (seq[k][i] - 'A');
            22             for(int k = n - 1; k >= 1; k--)
            23                 if((heap[k] = heap[2 * k] & heap[2 * k + 1]) == 0)
            24                 {
            25                     cost++;
            26                     heap[k] = heap[2 * k] | heap[2 * k + 1];
            27                 }
            28             char c = 'A';
            29             while(heap[1>>= 1)
            30                 c++;
            31             cout << c;
            32         }
            33         cout << ' ' << cost << endl;
            34     }
            35     
            36     return 0;
            37 }
            38 
            九九精品99久久久香蕉| 亚洲精品国产综合久久一线| 国产精品18久久久久久vr| 精品熟女少妇av免费久久| 999久久久免费国产精品播放| 久久se精品一区精品二区国产| 国产成人久久精品一区二区三区| 国产精品久久久久久影院| 日本精品久久久久影院日本 | 久久免费视频1| 国产成人久久精品激情| 久久最新免费视频| 九九久久自然熟的香蕉图片| 亚洲欧美国产精品专区久久| 久久国产精品成人免费| 亚洲欧美成人综合久久久| 精品视频久久久久| 久久99精品国产99久久6男男| 日韩va亚洲va欧美va久久| 青青草国产成人久久91网| 亚洲AV无码久久寂寞少妇| 久久亚洲精品国产精品婷婷| 久久久久久久国产免费看| 国内精品久久久久久久亚洲| AV狠狠色丁香婷婷综合久久| 亚洲人成网亚洲欧洲无码久久 | 欧美精品九九99久久在观看| 亚洲国产精品久久久久久| 久久亚洲AV成人无码电影| 久久久久久久精品成人热色戒 | 国内精品久久久久久久97牛牛| 久久久久久久97| 免费久久人人爽人人爽av| 亚洲精品国产第一综合99久久| 精品一久久香蕉国产线看播放| 99久久精品费精品国产| 亚洲国产成人久久综合碰碰动漫3d| MM131亚洲国产美女久久| 国产精品久久国产精品99盘 | 亚洲国产成人久久一区WWW| 久久99精品国产麻豆蜜芽|