A fast bit-parallel algorithm for matching extended regular expressions

This paper addresses the extended regular expression matching problem: given an extended regular expression (a regular expression with intersection and complement) r of length m and a text string x = a(1)... a(n) of length n, find all occurrences of substrings which match r. We will present a new bit-parallel pattern matching algorithm which runs in 0((mn(2) + ex(r)n(3))/W) time and 0((mn + ex(r)n(2))/W) space, where ex(r) is the number of extended operators (intersection and complement) occurring in r, and W is word-length of a computer. In addition, we actually implement the proposed algorithm and evaluate the performance.