求解Web集群文档分布的混沌搜索算法

Chaotic Searching Algorithm for Solving Document Distribution in Web Cluster Server

提出Web集群文档分布方案,用M/G/1/K PS排队模型对服务器进行建模,将文档分布问题转化为0-1整数规划问题,然后求解该规划问题。针对该类0-1整数规划问题,给出一种基于混沌搜索的求解算法,该算法让多个独立的混沌变量在其各自的轨道中搜索,使得对应生成的0-1矩阵能遍历任意一种可能的分布,从而能搜索到全局最优解。设计一种基于贪婪思想的文档分布算法。测试表明,混沌搜索算法能找到全局最优解,优于传统的贪婪算法。

Proposed document distribution strategy uses M/G/1/K PS queue to model Web server, and translates the document distribution problem into a 0-1 integer programming problem, and then solves such programming problem. Aiming at such kind of 0-1 integer programming problem, a chaotic searching algorithm is proposed to solve it. The algorithm lets many isolated chaotic variables search in their own tracks, so the corresponding 0-1 matrix built by these chaotic variables can experience every possible distribution, thereby it can find the global optimal solution. Considering the appearance of new Web pages, a greed algorithm of document distribution is also designed. Tests show that the chaotic searching algorithm can find the global optimal solution, and exceeds the traditional greed algorithm.