机构地区: 中国科学技术大学计算机科学与技术学院
出 处: 《软件学报》 2000年第6期771-778,共8页
摘 要: 并行串匹配算法的研究大都集中在 PRAM(parallel random access machine)模型上 ,其他更为实际的模型上的并行串匹配算法的研究相对要薄弱得多 .该文采用将最优串行算法并行化的技术 ,利用模式串的周期性质 ,巧妙地将改进的 KMP(Knuth- Morris- Pratt)算法并行化 ,提出了一个简便、高效且具有良好可扩放性的分布式串匹配算法 ,其计算复杂度为 O(n/ p+m) ,通信复杂度为 O(ulogp) ,其中 n为文本串长 ,m为模式串长 ,u为模式串最小周期长 ,p为处理器数 . Parallel string matching algorithms are mainly based on PRAM (parallel random access machine) computation model, while the research on parallel string matching algorithm for other more realistic models is very limited. In this paper, the authors present an efficient and scalable distributed string matching algorithm is presented by parallelizing the improved KMP (Knuth Morris Pratt) algorithm and making use of the pattern period. Its computation complexity is O(n/p+m) and communication time is O(u log p), where n is the length of text, m the length of pattern, p the number of processors and u the period length of pattern.
领 域: [自动化与计算机技术] [自动化与计算机技术]