Parallel Algorithms for Longest Common Subsequence Problem andAnt Colony Optimization Algorithms for Multiple-Longest Common Subsequence Problem

• Main Authors: Tse-Fu Chen, 陳則輔
• Other Authors: Hsiung-Chien Hsu
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/32537513777839733901

碩士 === 銘傳大學 === 資訊工程學系碩士班 === 92 === The longest common subsequence (LCS) problem has attracted many researchers’ attention since the ‘70s. Many sequential or parallel LCS algorithms can be found in the literature even since. We propose new parallel algorithms for the longest common subsequence (LCS) problem in the coarse grained multicomputers (CGM) model, and some ant colony optimization (ACO) algorithms for finding the longest common subsequence among multiple sequences (mLCS) in the thesis. For the LCS problem, we empirically study the performances of our and other CGM parallel algorithms in the literature on a PC cluster. To report the absolute speedup, Tsequential/Tparallel(p), the ratio of time needed by running the best sequential algorithm over the time needed by the parallel algorithm running on p processors, we also implement and test some elegant sequential algorithms for LCS. For the mLCS problem, we testify our ACO algorithms against some proposed approximation algorithms. Besides the parallel implementation of ACO is realized on a PC cluster. Experimental results about different kinds of input problem sets, such as DNA, protein or random sequences, demonstrate the efficiency and scalability of our CGM parallel algorithms for LCS and the ACO algorithms for mLCS.