A fast algorithm for multi-string matching based on automata optimization

Multi-string matching requires to handle massive amount of data in pattern recognition, intrusion detection, and biological sequence analysis applications. This paper proposes a new algorithm to construct an optimal automation to achieve fast string matching. The algorithm consists of five steps: sorting, forming subtrees, encoding all subtrees, similarity checking, and completing all transitions. The algorithmic complexity is proven O(umk), where u is the number of the symbols in the alphabet set and m and k are the average length and the number of strings being matched. We report analytical results on the matching complexity. These results prove the efficiency and effectiveness of the optimized automata generated for fast matching of multiple strings.