2012 Fiscal Year Final Research Report

An efficient POSIX regular expression matching via Glushkov automata with augmented transitions

Research Project

Project/Area Number	22500019
Research Category	Grant-in-Aid for Scientific Research (C)
Allocation Type	Single-year Grants
Section	一般
Research Field	Fundamental theory of informatics
Research Institution	Chubu University
Principal Investigator	OKUI Satoshi 中部大学, 工学部, 准教授 (00283515)
Co-Investigator(Kenkyū-buntansha)	SUZUKI Taro 会津大学, コンピュータ理工学部, 准教授 (90272179)
Project Period (FY)	2010 – 2012
Keywords	正規表現 / POSIX / パターン照合 / オートマトン
Research Abstract	Regular expression matching is crucial for many applications such as text processing. Although POSIX 1003.2 standard requires (sub-) matching to follow the leftmost-longest rules, almost none of existing implementations, which rely on backtracking, are responsible to the requirement; they follow the greedy semantics, an alternative way more suited for backtracking, instead. This study has offered, based on a slight extension of Glushkov automata (aka, position automata), a new and more efficient matching algorithm that accommodates POSIX regular expression matching. We have given its rigorous correctness proof, and the exact computational cost at worst case. Our experimental implementation has shown significant improvements on efficiency in some practical cases.