Abstract

We propose using proof plans to implement expression normalizers in automatic theorem proving. We outline some general-purpose proof plans and show how these can be combined in various ways to yield some standard normalizers. We claim that using proof plans facilitates the flexible application of these normalizers so that they can interact with the theorem prover in which they are embedded. We intend to extend this technique to decision procedures. 1 Introduction In [Boyer & Moore 88], Boyer and Moore investigate a case study in the use of decision procedures in automated deduction, namely a decision procedure for linear arithmetic. Their conclusions were as follows. ffl Such decision procedures have a vital role to play in reducing the combinatorial explosion, but ffl they cannot be treated as black boxes. In practice, few sub-goals exactly fit the requirements of a decision procedure, but many almost do. In these cases it is necessary to augment the decision procedure with ad...

Citations