带权选择活动.

Author: shengshijun

README.md

@@ -29,6 +29,10 @@ algorithm
 4.  加权有向无环图中最短路径和最长路径
 
 ###幂乘：算法复杂度是O(lgn)
+##贪心算法
+1.  活动选择问题
+2.  顶点作色问题
+3.  带权活动选择问题(其实就是一个调度问题）
 
 ###斐波那契树
 1.  使用循环实现的算法o(n)

dynamic/task/__init__.py

@@ -0,0 +1,3 @@
+#!/usr/bin/env python
+# -*- coding:UTF-8
+__author__ = 'shenshijun'

dynamic/task/max_weighted_action.py

@@ -0,0 +1,82 @@
+#!/usr/bin/env python
+# -*- coding:UTF-8
+__author__ = 'shenshijun'
+"""
+动态规划实现活动调度的问题，每个活动都有开始和结束时间，并且有一个权，要求得使得权和最大的活动集合
+"""
+
+
+class Action(object):
+    """
+    活动对象
+    """
+
+    def __init__(self, start_time, end_time, weight):
+        self.start_time = start_time
+        self.end_time = end_time
+        self.weight = weight
+
+    def __cmp__(self, other):
+        pass
+
+    def __eq__(self, other):
+        return self.start_time == other.start_time and self.end_time == other.end_time
+
+    def __hash__(self):
+        return hash(hash(self.start_time) * hash(self.end_time))
+
+    def __str__(self):
+        return "".join(["Action(start_time=", str(self.start_time), ",ent_time=", str(self.end_time), ",weight=",
+                        str(self.weight), ")"])
+
+
+__value_dict = {}
+__actions_dict = {}
+
+
+def max_weighted_actions(action_list):
+    """
+    假设Action是按照end_time排序的。因此在算法里面就不排序了。
+    :param action_list:
+    :return:
+    """
+    global __actions_dict
+    global __value_dict
+    action_len = len(action_list)
+    # 循环退出条件
+    if action_len is 0:
+        return 0
+
+    cur_max_weight = 0
+    for k in xrange(action_len):
+        left_max_weight = 0
+        if k > 0:
+            if (action_list[0], action_list[k]) not in __value_dict:
+                __value_dict[(action_list[0], action_list[k])] = max_weighted_actions(
+                    filter(lambda action: action.end_time <= action_list[k].start_time, action_list[:k]))
+            left_max_weight = __value_dict[(action_list[0], action_list[k])]
+
+        right_max_weight = 0
+        if k < action_len - 1:
+            if (action_list[k], action_list[-1]) not in __value_dict:
+                __value_dict[(action_list[k], action_list[-1])] = max_weighted_actions(
+                    filter(lambda action: action.start_time >= action_list[k].end_time, action_list[k + 1:]))
+            right_max_weight = __value_dict[(action_list[k], action_list[-1])]
+
+        cur_weight = left_max_weight + right_max_weight + action_list[k].weight
+        if cur_max_weight < cur_weight:
+            cur_max_weight = cur_weight
+
+        __value_dict[(action_list[0], action_list[-1])] = cur_max_weight
+    return __value_dict[(action_list[0], action_list[-1])]
+
+
+def main():
+    action_list = [Action(1, 4, 2), Action(3, 5, 1), Action(0, 6, 1), Action(5, 7, 2), Action(3, 9, 1), Action(5, 9, 1),
+                   Action(6, 10, 1), Action(8, 11, 2), Action(8, 12, 1), Action(2, 14, 1), Action(2, 16, 2)]
+    print(max_weighted_actions(action_list))
+
+
+if __name__ == "__main__":
+    main()
+

greedy/__init__.py

@@ -0,0 +1,3 @@
+#!/usr/bin/env python
+# -*- coding:UTF-8
+__author__ = 'shenshijun'