Abstract: The rough set theory is a new method to deal with incomplete and uncertain data. The procedure for the theory to process data is very similar to that for the train dispatcher to make decision to adjust the conflicting trains. Both of them extract the decision-making rules from a large number of experienced data with which then the new decision are gained. Through analyzing the work of train operation adjustment and the decision-making method applied by dispatchers in detail, a model on the Rough Set is developed, and the steps for obtaining and expressing the train operation adjustment knowledge are put forward. Furthermore, the reduction algorithm of decision-making rules of train operation adjustment is also improved. All these make the conception for the method clearer and the solution making steps perfect. Finally, by mining the data of train operation from DMIS, the practicality of the algorithm method above is validated, and because the effecting extents of conditional attributes to the decision-making could be evaluated objectively from the extracted rules, the method discussed could provide valuable references for the train dispatchers to make new effective adjustment measures, and consummate train operation adjusting system on Rough Set Theory.

Key words: railway transportation, artificial intelligence, train operation adjustment, decision-making based on Rough Set, data mining

摘要： 粗糙集是一种处理不完备、不精确数据的新方法，其数据处理过程与行车调度员进行列车运行调整的决策过程非常相似，都是从已有的大量经验数据中提取出决策规则并用于指导新的决策判断．通过对行调人员运行调整和决策的详细剖析，构建了列车运行调整的粗糙集模型，提出了基于粗糙集的列车运行调整系统知识的获取与表达的方法，改进了运行调整的粗糙集决策规则的约简算法，思路更清晰，步骤更完善；最后通过对DMIS系统中部分列车运行数据的挖掘，证明了算法的实用性，从提取出的规则能直观地判断出各条件属性对决策的影响程度，为行调做出新的决策提供了有价值的参考；形成了完整的粗糙集列车运行调整体系．

关键词: 铁路运输, 人工智能, 运行调整, 粗糙决策, 数据挖掘