Optimizing Bottom Up Query Evaluation For Deductive Databases

Optimizing Bottom-up Query Evaluation for Deductive Databases

On the other hand, for many programs, Prolog is arbitrarily slower than bottom-up evaluation. Our optimization techniques are of importance in evaluating programs that generate facts concerning variables. In the second part of the thesis, we develop optimizations related to the use of aggregate operations such as min or max. We show how to view several such operations as 'selections', and how to propagate these selections into programs. We demonstrate the power and utility of the optimization techniques, using programs for problems such as computing shortest paths and critical paths."

Query Optimization in Deductive and Relational Databases

Computer Sciences Technical Report