A hybrid parallel evaluation model for logic-based intelligent systems

Abstract

We present a hybrid model to speed up the evaluation of a logic based intelligent system. A logic based system is first applied by a data dependency analysis technique which can find all the mode combinations that exist within clauses of a knowledge base. The mode information is used to support a novel hybrid parallel evaluation model, which combines both top down and bottom up evaluation strategies. This model can preserve maximum parallelism while guaranteeing to generate all the solutions of a logic based knowledge base without backtracking. The overall parallel execution behavior of the logic based system can thus be improved by reducing the total number of nodes searched in the tree, the total processes needed to be generated and the total communication channels needed in the search process. A simulator has been implemented to analyze the execution behavior of the new model. Experiments show significant improvement under most situations.