Study On Optimization Of Multi-Objective Maneuvering Of Trains Based On Nsga-?

With the rapid development of urban rail transit,the passenger flow at subway stations is increasing.Metro trains play an important role in rail transit.As an important transportation tool in the city,the operation of subway trains must strictly comply with various indicators,so as to ensure that the operating efficiency can be improved under the premise of safe driving and passenger travel requirements can be met.However,in the actual operation of the subway,it will be affected by lines,speeds and other vehicles,which will often increase the operation energy consumption,time,and reduce passenger comfort,thereby increasing the operating costs.Therefore,the multi-objective research on the train operation process is particularly important.This paper takes the train's running energy consumption,time,and passenger comfort as the optimization goals.Traffic safety and precise parking are the constraints.The NSGA-? algorithm(Non-dominated Sorting Genetic Algorithm ?)is used to track the train's multi-target operation process and trains process for optimization research.(1)Study the train's operation process.Firstly,the principle of the automatic train control system and the indicators of train operation are analyzed.Then,according to the running process of the train,the force analysis was carried out.Combined with the train's own characteristics and the line conditions,the train's multi-mass movement model was established.(2)Study the optimization problem of multi-objective genetic algorithm.Traditional multi-objective optimization methods often turn multi-objective problems into single-objective problems,resulting in the loss of many optimal solutions.The NSGA-II algorithm is prone to the situation of uneven distribution of the solution during optimization,thereby reducing the diversity of solution sets.Aiming at this problem,this paper introduces a reference point-based NSGA-? algorithm to enhance the optimization ability of the algorithm and improve the diversity and distribution uniformity of the solution set.Using the standard test function to compare the NSGA-? algorithm based on the reference point with NSGA-II,it is verified that the NSGA-? algorithm has better optimization performance and lays the foundation for the multi-objective optimization of the train.(3)Research on multi-objective maneuver optimization of trains.For the actual running status of the train,the golden section genetic algorithm is used to optimize the train's energy consumption,time,and passenger comfort,and various optimization indicators are obtained.Then considering the running energy consumption,time and passenger comfort of the train as the optimization goal,the driving safety,the parking precision and the working conditions are constraints,and an optimized multi-objective train operation model is established.For the problem of multi-objective optimization for a constrained train,the concept of constraint violation is introduced to expand NSGA-? so as to reduce the generation of infeasible solutions in the optimization and improve the quality and diversity of the solution set and adopt the extended NSGA-? algorithm.The multi-objective maneuver optimization model of the train is solved.Combining with the actual subway lines,the single-objective optimization and multi-objective optimization performance of trains were compared by simulation to verify the necessity of multi-objective maneuver optimization of trains.(4)Study the train multi-objective tracking optimization problem.On the basis of multi-objective maneuver optimization of trains,considering the influence of multiple cars comprehensively,the inter-vehicle tracking interval time is added to the constraints of trains,and a train tracking optimization model is established.According to the track running process of the train,three different tracking scenarios were established and corresponding tracking strategies were established.Combined with its strategy,the NSGA-? algorithm was used to solve the train tracking model.Combining with the actual subway lines,it is verified through experimental simulation that under the optimization of the algorithm,the various indexes of the train reach a good level during the tracking process,which proves the feasibility and effectiveness of the algorithm.