Study On Knowledge Reduction Method Of Energy-saving Oriented Rail Transit Metro Diagram

With the people’s travel demand increasing accordingly,Urban rail transit has been widely applied due to the characteristics of large passenger volume,fast and punctual.However,the proportion of urban rail transit in urban traffic is increasing,and its energy consumption is also increasing.When faced with an emergency,it takes a long time to compile a rail transit metro diagram by using the simulation system,which can’t meet the needs of rail transit enterprises and passengers.Therefore,in order to quickly retrieve an energy-saving oriented“optimal metro diagram”,case-based reasoning（CBR）is introduced into the design of rail transit metro diagram case base.With the increasing demand of intelligent system,the factors affecting the decision-making target of knowledge base are increasing,which makes the problem of knowledge redundancy of case base more and more serious.It not only reduces the decision efficiency of case-based reasoning system,but also increases the time of case retrieval.Therefore,in the premise of the case-based reasoning system decision-making unchanged,putted forward by rail transit train diagrams for knowledge reduction and improve the accuracy of case retrieval results,shorten the time of the case retrieval,is the research target of this article.Based on the analysis of domestic and overseas advance,this paper makes knowledge reduction on the case base of rail transit metro diagram from two aspects: attribute reduction and value reduction of case base.In view of the traditional rough set method can only deal with discrete numerical problem,the attribute reduction method based on fuzzy rough sets is introduced.For dynamic fuzzy clustering method is improved,and a method for determining the parameters determined by F-measure clustering threshold.Based on the original data distribution characteristics of the case database,the accuracy of attribute reduction results is improved.In the process of attribute reduction,the fuzzy similarity matrix is established by using the Euclidean distance and the included angle cosine method,and the accuracy by the two similarity algorithms is compared under the F-statistic,XB and F-measure validity respectively.The attribute reduction method of fuzzy rough set based on the combination of Euclidean distance and F value has an accuracy of 93.33%.The heuristic value reduction algorithm is used to reduce decision table of the rail transit data set.A case base knowledge acquisition model based on rough set is constructed to extract rules from the case base of rail transit metro diagram.On that basis,the rule representation of the case base information is carried out,and the rail transit metro diagram case base integration model based on rule and CBR is constructed.The case-based reasoning model for large-scale data set of metro diagram retrieval provides a new method.Respectively in the energy consumption and the retrieval time to verify the effectiveness with above method of knowledge reduction for case retrieval.In the design of urban rail transit metro diagram CBR system,a comprehensive weight determination method based on fuzzy rough set and AHP is proposed,and a genetic algorithm model is used to optimize the attribute weights.Based on nearest neighbor with entropy weight similarity calculation method for mixed attribute,the accuracy and case discrimination of the similarity results are improved.Experiments show that the case retrieval system with reduced knowledge can improve the precision and retrieval time significantly.