Research On Multi-objective Urban Bus Dispatching Optimization Method

The acceleration of the urbanization process has caused too many private vehicles to occupy originally few road resources,which has continuously restricted urban development.To resolve these contradictions,the public transportation model has been vigorously promoted.However,bus line planning and infrastructure construction have long periods,slow results and huge costs,and bus scheduling,which is one of the core tasks of bus operations,is flexible and efficient.However,with the advancement of urbanization,many new features have emerged in urban transportation,and many traditional scheduling methods and theoretical applications are inefficient.Therefore,how to formulate bus scheduling strategies more efficiently,scientifically and accurately has a positive impact on the actual operation of buses.Most of the current bus dispatch optimization strategies assume that the vehicle maintains a constant speed within the operating interval,the optimization goal is single,the lack of dispatch control strategies,the lack of real-time in practical applications,it is difficult to explore accurate bus operation rules,which is not conducive to the development of scientific and effective bus departures interval.In this regard,this article introduces new data processing methods,improves the dispatch model,and increases the adaptive adjustment of the departure interval,thereby providing a reference for the optimization of bus dispatch.The main research contents of this article are as follows:1.Based on the advanced theories and applications of bus dispatching in various countries,this article first discusses the basic mode and influencing factors of bus dispatching under the current development situation.By comparing the performance of traditional bus dispatch optimization methods with intelligent algorithms such as evolutionary algorithm,swarm intelligence algorithm,simulated annealing algorithm,etc.,NSGA2 algorithm was determined as the model solving algorithm.Then,data mining method is used to clean and fuse the acquired data,and the travel characteristics are described from the time and space dimensions.Finally,the Gaussian hybrid clustering algorithm is used to describe the hidden rules of travel characteristics from a deep perspective.Using the above method,the travel data of passengers at various stations along the bus line are combed and analyzed to provide correct data reference and optimization premise for bus dispatch.2.In order to avoid the waste of time cost and economic cost,a three-objective optimization function including passenger travel time cost,bus operation cost,ride comfort,etc.was constructed to calculate a more scientific,reasonable and humanized scheduling schedule.Although the NSGA2 algorithm used has high calculation efficiency,high individual good rate,and can calculate a relatively uniform Pareto frontier,in practical applications,it is found that the NSAG2 algorithm is easy to cause non-dominated solution loss and easy to fall into local optimal.Therefore,with the help of simulated annealing,the jump probability is added to the individual selection strategy of the NSGA2 algorithm to obtain the SA-NSGA2 algorithm.And through the test function,it is proved that the SA-NSGA2 algorithm can obtain a better quality and more evenly distributed solution set.3.Finally,through an example analysis of the bus operation data of Shenzhen No.113 bus,it is proved that the introduced data mining method can accurately and quickly show the travel characteristics of bus passengers in the time domain and the space domain.Moreover,the scheduling schedule obtained by the simulation experiment can not only improve the passenger's travel efficiency and ride happiness,but also reduce the cost of bus operation,which proves the effectiveness and advancedness of the scheduling optimization method adopted in this paper.