Systolic-based parallel architecture for the longest common subsequences problem

In this paper we design a new and efficient systolic architecture for the longest common subsequences problem which is, given two finite strings on any alphabet, to recover a subsequence of maximal length of both strings. A natural extension to this problem is to determine the set of all longest common subsequences of the two given strings. First, we present a modular linear time algorithm on an input/output bounded and fault-tolerant semi-mesh systolic structure for the longest common subsequence problem. Then, we extend this algorithm to the set of all longest common subsequences problem. (C) 1998 Elsevier Science B.V. All rights reserved.