| As the backbone of modern urban transportation,urban rail transit plays an important role in promoting the sustainable development of urban economy,alleviating traffic contradictions,saving energy and protecting the environmental,and solving the problem of stable employment.With the rapid development and large-scale application of various new technologies such as the Internet of Things,Big Data,and Artificial Intelligence.Fully Automatic Operation(FAO)has become the development direction of a new generation of urban rail transit.Compared with the manual driving mode of traditional urban rail transit,the FAO system achieves a fully automated,unmanned operation mode,and has higher train operation control accuracy.During the operation of a fully automatic train,it will be affected by multiple influences such as the line environment,train speed,and vehicle scheduling.The pros and cons of its tracking operation control strategy will directly affect the safety,energy saving,punctuality,parking accuracy and passenger comfort of the train.Therefore,the study of efficient fully automatic train tracking operation control strategy has important practical significance for the safety and efficiency of train operation.The main research contents of this paper are as follow:(1)According to the operation characteristics of fully automatic train operation of urban rail transit,combined with the current research status of train tracking operation optimization at home and abroad,the entire train is simplified into a single point model for force analysis.Summarize the operating conditions and driving strategies of fully automatic trains,analyze the performance indicators that need to be optimized,and establish mathematical models corresponding to the four optimization indicators.(2)Multi-objective optimization algorithms and improvement strategies are studied.Combining the theory of multi-objective optimization,it analyzes the shortcomings of the Non-dominated Sorting Genetic Algorithm Ⅲ(NSGA-Ⅲ)based on reference points and proposes an improvement strategy,leading to the Elite-Opposition-based Learn NSGA-Ⅲ(EOL-NSGA-Ⅲ)algorithm.A series of different multi-objective test functions are used to compare and prove the optimization effects of the algorithm before and after the improvement,which proves that the EOL-NSGA-Ⅲ algorithm has better convergence,diversity and solution performance.(3)Comprehensively consider multiple performance index requirements and constraints such as speed limit,running time and working condition conversion to establish a multi-objective train operation optimization model.Using the actual fully automatic route and train information,the model is solved based on NSGA-Ⅲ algorithm and EOL-NSGA-Ⅲ algorithm,and the corresponding optimization results are obtained.The simulation verifies that the optimization effect of the EOL-NSGA-Ⅲ algorithm is superior.(4)According to the train tracking operation process of two trains on the same route,it is divided into different tracking scenarios and corresponding optimization strategies are proposed.Combining multiple optimization index models and tracking constraints,an optimization model for train tracking operation is established.Apply actual fully automatic route and train tracking operation data,the EOL-NSGA-Ⅲ algorithm is used to simulate and analyze the optimization process of train tracking operation in different cases.The simulation results confirmed that,according to the proposed tracking operation optimization strategy,the train operation speed curve optimized by the EOL-NSGA-Ⅲ algorithm was used to achieve train safety,energy-saving,punctual and smooth operation. |
