数与图的完美结合—浅析差分约束系统.doc

数与图的完美结合 -------浅析差分约束系统 华中师大一附中 冯威 [摘要] 在面对多种多样的问题时，我们经常会碰到这样的情况：往往我们能够根据题目题面意思来建立一些简单的模型，但却面对这些模型无从下手。这时我们应该意识到，也许能够将这种模型与其他的模型之间搭起一座桥梁，使我们能够用更简单直接的方式解决它。这里我们介绍一种方法，它很好地将某些特殊的不等式组与图相联结，让复杂的问题简单化，将难处理的问题用我们所熟知的方法去解决，它便是差分约束系统。这里我们着重介绍差分约束系统的原理和其需要掌握的bellman-ford算法。然后通过zju1508和zju1420两道题目解析差分约束系统在信息学题目中的应用，并逐渐归纳解决这类问题的思考方向。 [目录] ◆ 关键字 ······························································································· 【2】 ◆ Bellman-ford算法 ············································································ 【2】 ◇ 算法简单介绍 ················································································· 【2】 ◇ 算法具体流程 ················································································ 【2】 ◇ 例题一ZJU2008 ·············································································· 【4】 ◆ 差分约束系统 ···················································································· 【5】 ◇ 例题二ZJU1508 ·············································································· 【5】 ◇ 线性程序设计 ················································································· 【7】 ◇ 差分约束系统 ················································································· 【7】 ◇ 例题三 ZJU1420 ············································································ 【8】 ◆ 结语 ··································································································· 【9】 ◆ 附录 ··································································································· 【9】 [关键字] 差分约束系统、不等式、单元最短路径、转化 [正文] 在分析差分约束系统之前，我们首先介绍一个解决单元最短路径问题的Bellman Ford算法，它的应用十分广泛，在差分约束系统中更充当着重要的角色。 Bellman-ford 算法 算法简单介绍 这个算法能在更一般的情况下解决最短路的问题。何谓一般，一般在该算法下边的权值可以为负，可以运用该算法求有向图的单元最长路径或者最短路径。我们这里仅以最短路径为例。 Bellman ford 类似于Dijkstra算法，对每一个节点vV，逐步减小从起点s到终点v最短路的估计量dist[v]直到其达到真正的最短路径值mindist[v]。Bellman-ford算法同时返回一个布尔值，如果不存在从源结点可达的负权回路，算法返回布尔值TRUE，反之返回FALSE。 算法具体流程 枚举每条边(u,v)E（G）。 对枚举到的边进行一次更新操作。 回到步骤1，此过程重复n-1次，以确定没有更可以优化的情况。 枚举每条边（u，v）若