In this dissertation, we reconsider some general SAT techniques under the original DPLL framework. These SAT techniques include unit propagation, variable ordering, value ordering, linear/nonlinear programming, and constraint strengthening. We also study performance differences of SAT techniques under different clause to variable ratios and under different SAT structures converted from underlying real-world applications.Table 5.5: The number of instances each method can solve for the IPC3-4 STRIPS domains. N is the total number of problem instances in each domain. We highlight in box the best one among four algorithmic combinations. Table 5.6: Theanbsp;...
|Title||:||Searching for Optimization Through Satisfiability|
|Publisher||:||ProQuest - 2008|