基于实例的客运站股道运用实时决策推理方法

Method of case-based real-time decision deduction of track utilization in railway passenger stations

客运站股道运用实时决策问题是一个半结构化问题,实时决策推理研究为解决特殊情况下铁路客运站实时决策问题提供理论基础,具有十分重要的现实意义。采用物元理论,构建基于实例的实时决策推理物元模型,将实时决策物元分为列车物元、方案物元和推理物元,结合车站调度员解决特殊情况下股道运用实时决策的思维过程,采用优度评价法和基于双空间模型的归纳学习法,设计基于实例的实时决策推理算法和自学规则算法,提出铁路客运站股道运用实时决策推理方法,用于制订特殊情况下的实时决策方案。实例分析结果表明,所提出的模型与算法能有效解决特殊情况下客运站股道运用实时决策问题,提高了股道运用实时决策的智能化水平。

Real-time decision problem of track utilization in railway passenger stations is a type semi-structural problem and its deduction can provide theoretical basis for solving real-time decision problem with special conditions,which has very important practical meaning.Adopting matter element theory,this paper constructed case-based real-time decision model and divided real-time decision matter element into train matter element,plan and deduction ones.Then,it combined with the thinking process of dispatchers for solving real-time decision problem with special conditions,adopted superiority evaluation method and learning method based on double-space model,designed case-based real-time decision deduction algorithm and rule-self learning algorithm,and put forward methods of case-based real-time decision deduction of track utilization,to make real-time decision plans with special conditions.The instance results indicate that the model and algorithms can solve the real-time decision problem with special conditions effectively and the intelligence level of real-time decision of track utilization can also improve greatly.