TY - GEN

T1 - Compacting a dynamic edit distance table by RLE compression

AU - Hyyrö, Heikki

AU - Inenaga, Shunsuke

PY - 2016

Y1 - 2016

N2 - Kim and Park [A dynamic edit distance table, J. Disc. Algo., 2:302–312, 2004] proposed a method (KP) based on a “dynamic edit distance table” that allows one to efficiently maintain edit distance information between two strings A of length m and B of length n when the strings can be modified by single-character edits to their left or right ends. This type of computation is useful e.g. in cyclic string comparison. KP uses linear time, O(m + n), to update the distance representation after each single edit. As noted in a recent extension of KP by Hyyrö et al. [Incremental string comparison, J. Disc. Algo., 34:2-17, 2015], a practical bottleneck is that the method needs Θ(mn) space to store a representation of a complete m×n edit distance table. In this paper we take the first steps towards reducing the space usage by RLE compressing A and B. Let M and N be the lengths of RLE compressed versions of A and B, respectively. We propose how to store the edit distance table using Θ(mN + Mn) space while maintaining the same time complexity as the original method that does not use compression.

AB - Kim and Park [A dynamic edit distance table, J. Disc. Algo., 2:302–312, 2004] proposed a method (KP) based on a “dynamic edit distance table” that allows one to efficiently maintain edit distance information between two strings A of length m and B of length n when the strings can be modified by single-character edits to their left or right ends. This type of computation is useful e.g. in cyclic string comparison. KP uses linear time, O(m + n), to update the distance representation after each single edit. As noted in a recent extension of KP by Hyyrö et al. [Incremental string comparison, J. Disc. Algo., 34:2-17, 2015], a practical bottleneck is that the method needs Θ(mn) space to store a representation of a complete m×n edit distance table. In this paper we take the first steps towards reducing the space usage by RLE compressing A and B. Let M and N be the lengths of RLE compressed versions of A and B, respectively. We propose how to store the edit distance table using Θ(mN + Mn) space while maintaining the same time complexity as the original method that does not use compression.

UR - http://www.scopus.com/inward/record.url?scp=84956657208&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84956657208&partnerID=8YFLogxK

U2 - 10.1007/978-3-662-49192-8_25

DO - 10.1007/978-3-662-49192-8_25

M3 - Conference contribution

AN - SCOPUS:84956657208

SN - 9783662491911

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 302

EP - 313

BT - SOFSEM 2016

A2 - Freivalds, Rūsiņš Mārtiņš

A2 - Engels, Gregor

A2 - Catania, Barbara

PB - Springer Verlag

T2 - 42nd International Conference on Current Trends in Theory and Practice of Computer Science, SOFSEM 2016

Y2 - 23 January 2016 through 28 January 2016

ER -