New implementation technique for model generation theorem provers to solve constraint satisfaction problems

A new implementation technique for model-generation theorem provers (MGTP) is presented, which is particularly efficient in solving constraint satisfaction problems. We solved the so-called multi-environment problem by using restoration lists for destructively modified data structures. Furthermore, we introduced a new mechanism called Activation-cell which dramatically reduces the complexity of conjunctive matching, subsumption tests, and conflict tests in the proving process of MGTP. Combined with other implementation techniques including term-indexing and clause-indexing, the new MGTP written in Java shows remarkable performance compared to the previous implementations.