计算本体的概念层次结构是描述逻辑(Descriptic Logic,DL)推理机提供的一个重要功能.本文研究了支持DL推理的Pellet1.4(简称Pellet)推理机的概念分类算法,分析了影响算法性能的主要因素,在此基础上给出了一个基于全面标志机制的优化改进算法。算法实验结果表明,优化算法可以有效的减少调用DL- ABOX执行一致性检测次数,提高本体概念分类性能。 The conceptual hierarchy of computational ontology is an important function provided by descriptive logic (DL) inference engine.In this paper, the conceptual classification algorithm of Pellet1.4 (referred to as Pellet) inference engine supporting DL inference is studied, Based on this, an optimization and improvement algorithm based on comprehensive marking mechanism is given. The experimental results show that the optimization algorithm can effectively reduce the number of DL-ABOX calls and improve the ontology conceptual classification performance.