High-speed Train Ride Comfort Evaluation Method Based On IFOA-BPNN

With the continuous optimization and improvement of high-speed railway network,as well as the improvement of high-speed train speed and safety in the current operation process of high-speed railway,high-speed train has become the preferred means of transportation for people to travel,while passengers put forward higher requirements for comfort.The evaluation of high-speed train ride comfort is mainly the comprehensive evaluation of passengers to train service and riding environment.According to a large number of existing studies on high-speed train ride comfort,it is found that the mathematical modeling process of high-speed train ride comfort evaluation is very complicated and mixed,and usually the evaluation results will be one-sided and speculative emotional color,unable to directly reflect the intuitive feeling of passengers riding high-speed train.Based on this,a comfort evaluation model based on IFOA-BPNN is constructed in this paper by combining the improved Drosophila optimization algorithm and back-propagation neural network theories.The selected comfort evaluation index is taken as the network input,and the results obtained from fuzzy comprehensive evaluation are taken as the output of the network model.An effective evaluation model is constantly trained and optimized.The complicated calculation process of traditional evaluation method is avoided,and the passenger satisfaction of high-speed train comfort is obtained.For the problem of multiple excellent individuals being simultaneously deleted in Non-dominated Sorting Genetic algorithm-II(NSGA-II),a corresponding strategy improvement Algorithm was proposed to improve population diversity.At the same time,according to the evaluation results of passengers on the thermal environment of high-speed trains,both energy consumption and comfort were taken into account,the corresponding multi-objective optimization function was established,the constraint conditions were determined,and the improved NSGA-II algorithm was used to further optimize the parameters of the thermal environment air conditioning system of the carriage,so as to comprehensively improve the thermal environment comfort of the carriage.Firstly,aiming at the problem of weight determination in the comprehensive model of fuzzy evaluation of high-speed train comfort,Using FAHP(Fuzzy Analytic Hierarchy Process)and PCA(Principal Component Analysis)to determine the weight of evaluation index from subjective and objective aspects,Then,the weight of these two parts is optimized and fused to obtain the final weight result.Then,the final comfort evaluation result is obtained according to the fuzzy comprehensive evaluation method,which provides theoretical data support for the model in the following paper.Secondly,dissertation combines the improved Drosophila optimization algorithm with the back propagation neural network,and proposes the IFOA-BPNN model to comprehensively evaluate the comfort level.Aiming at the problem that traditional back-propagation neural networks are easy to fall into the local optimal solution,an improved Drosophila optimization algorithm is adopted to optimize the initial weights and thresholds of the algorithm.Firstly,cubic chaotic mapping was used to enhance the diversity of Drosophila population initialization,so as to solve the problem that the initial position of the population assigned to the individual initial position is not uniform due to the random selection of the function in the standard Drosophila algorithm,and the error may occur in the optimal solution.Secondly,in order to solve the problem that drosophila individual search step size is randomly set in traditional Drosophila algorithm,and its size affects its local search ability and convergence speed,this paper proposes a step-changing mechanism that dynamically ADAPTS to change individual search step size,so that the algorithm has a good global search ability in the early iteration stage and a strong local search ability in the late iteration stage.Finally,the problems existing in the traditional NSGA-II algorithm are analyzed and explored.On the one hand,when the binary tournament selection method selects population individuals,it cannot guarantee that the selected individuals perform better than the unselected ones.On the other hand,the crowding degree distance will cause the algorithm to delete and eliminate excellent individuals at the same time,resulting in uneven distribution and loss of individuals in the solution set space.An improved crowding degree distance judgment strategy is proposed,which makes the genetic selection process integrate the characteristics of target space and decision space,ensure the diversity of individuals in the population,and avoid the multi-objective single processing.Combined with the PDD-PMV model of thermal environment comfort,a multi-objective optimization function based on the comfort of high-speed train and the energy loss of air conditioning system in thermal environment is constructed,and the improved algorithm is used to solve it,and the most suitable solution can be selected in the solution space according to the specific requirements of the real situation.