Efficient method to generate minimal models by introducing splitting lemmas

An efficient method for minimal model generation is presented. The extended MGTP, called MM-MGTP, was implemented using the method so that every model it generates is guaranteed to be minimal. The method employs splitting assumption which is equivalent to Bry's complement splitting rule as a basic mechanism. Moreover, we introduce a new concept called splitting lemma which can avoid performing unnecessary tests on minimal models and prune branches leading to non-minimal models. Experimental results with MM-MGTP show remarkable speedup compared to Bry's MM-SATCHMO.