### Refine

#### Document Type

- Preprint (3) (remove)

#### Keywords

- resolution (2)
- Equality reasoning (1)
- Knuth-Bendix completion algorithm (1)
- compilation (1)
- narrowing (1)
- paramodulation (1)
- problem formulation (1)
- tactics (1)
- theorem proving (1)

#### Faculty / Organisational entity

- Fachbereich Informatik (3) (remove)

We transform a user-friendly formulation of aproblem to a machine-friendly one exploiting the variabilityof first-order logic to express facts. The usefulness of tacticsto improve the presentation is shown with several examples.In particular it is shown how tactical and resolution theoremproving can be combined.

We present a framework for the integration of the Knuth-Bendix completion algorithm with narrowing methods, compiled rewrite rules, and a heuristic difference reduction mechanism for paramodulation. The possibility of embedding theory unification algorithms into this framework is outlined. Results are presented and discussed for several examples of equality reasoning problems in the context of an actual implementation of an automated theorem proving system (the Mkrp-system) and a fast C implementation of the completion procedure. The Mkrp-system is based on the clause graph resolution procedure. The thesis shows the indispensibility of the constraining effects of completion and rewriting for equality reasoning in general and quantifies the amount of speed-up caused by various enhancements of the basic method. The simplicity of the superposition inference rule allows to construct an abstract machine for completion, which is presented together with computation times for a concrete implementation.