University of Houston

In normalization, traversing the term to find a match with a rule costs most of the time. This thesis presents the design, implementation and integration of a unifcation-based preprocessor into the Laboratory for Rapid Rewriting, LRR, and the current status of LRR. LRR consists of two interpreters: Smaran, which stores the history of all rule applications, and TGR, which stands for Term Graph Rewriter. We have improved the preprocessor and the DS-list significantly and efficiently integrated their latest versions into both components of LRR. The performance of the latest version of LRR on some benchmarks - both favorable and unfavorable - is presented and compared with other interpreters including Maude and Rascal.