The aim of this work is to provide a model for the dynamic implementation of finite automata for enhanced performance. Investigations have shown that hardcoded finite automata outperforms the traditional table-driven implementation up to some threshold. Moreover, the kind of string being recognized plays a major role in the overall processing speed of the string recognizer. Various experiments are depicted to show when the advantages of using hardcoding as basis for implementing finite automata (instead of using the classical table-driven approach) become manifest. The model, a dynamic algorithm that combines both hardcoding and table-driven is introduced.
|Tijdschrift||International Journal of Foundations of Computer Science|
|Nummer van het tijdschrift||6|
|Status||Gepubliceerd - 2005|