Regular Expression Simplification
This paper was my 3rd year undergraduate project, and a regular expression simplifier was written in Haskell with reasonably good performance.
Regular expressions are a widely used concept, which are commonly manipulated by first converting them to Finite State Automata. This conversion has a high computational complexity associated with it. The corresponding Finite State Automata to regular expression conversion algorithm produces far from minimal expressions.
This project reviews the existing theory of regular expressions, including current research and existing implementations of systems, focusing on the simplification aspect. A program is designed and implemented to attempt the specific problem of simplification, using conditional rewrite rules.
The resulting program is proved to always produce equivalent regular expressions, and to always terminate. A function is also given that measures the complexity of a regular expression. The program performs significantly better than any other solution available, in terms of simplification, but does not produce minimal expressions in every case.