铁路行车事故救援决策方法研究

Decision-Making Methods for Train Operation Accident Rescue

传统的铁路行车事故救援多采用人工方式给出救援方案,但事故受多方面因素的影响,救援人员很难及时的给出科学合理的救援方案。针对已有救援知识不完备、不系统的特点,提出规则推理(Rule-based Reasoning,RBR)和案例推理(Case-Based Reasoning,CBR)相结合的两级分层推理框架,给出了系统流程图,说明了RBR与CBR的具体实现方法,并将自组织特征映射网络(Self-Organizing Feature Map,SOFM)应用到事例检索中,有效地提高了检索的效率。仿真实验结果表明系统取得了良好的效果。克服了单一推理的缺点,实现了对救援理论和经验的复用,提高了系统的效率和综合推理能力,并使系统具有了学习能力。研究结果为进一步应用奠定了基础。

The traditional train operation accident rescue often depends on manual basis to get rescue schemes,but it is hard for the rescuers to arrive to reasonable schemes in time when considering all sorts of influencing factors.In view of the imperfection and unsystematic characteristic of rescue knowledge,this paper brings forward a two-level reasoning frame,which combines Rule-based Reasoning with Case-Based Reasoning,introduces system flow chart,illuminates the implementation method of RBR and CBR,and applies Self-Organizing Feature Map to case retrievals,which levels up the efficiency of retrieving.The simulation results show that the system functions well.This scheme overcomes the disadvantage of monotonic reasoning,actualizes the re-use of rescue theories and experience,improves the efficiency and integrative-reasoning ability of the system,and endows the system with the ability of study.The outcome of this research establishes the foundation for further application.