Research And Realization Of Bus Schedule Optimization Based On Big Data

At present,the bus dispatching company in my country mainly arranges bus schedules based on previous manual experience.Such scheduling is not only very cumbersome but also cannot improve bus operation efficiency.This thesis will use bus big data to optimize bus scheduling.The optimization of public transportation scheduling can adapt to people’s travel rules,dredge the flow of people,improve the efficiency of public transportation,and attract more and more people to choose public transportation.This thesis first analyzes the data source and preprocesses the passenger’s IC card data.Use these data to mine and extract the passenger flow of passengers on each route,and analyze the travel time distribution of passengers.According to the passenger flow of each station,the bus travel distance and the land nature of the drop-off station are used to calculate the drop-off probability and number of people at each station based on big data analysis.This article also analyzes and digitizes the influencing factors of passenger flow,and establishes a multiple linear regression model to predict the future passenger flow.This thesis analyzes the current public transportation enterprise dispatching system and establishes a mathematical model of public transportation scheduling.Based on the results of data mining analysis,this paper uses cross-sectional passenger flow data to constrain the model,and derives the spatial distribution of public transit passenger outflows.Then take the departure interval as the decision variable,and take the maximum departure interval,the minimum departure interval and the full load rate of the vehicle as the constraint conditions,establish the mathematical model of bus line scheduling,and take the minimum departure cost of the bus company and the minimum cost of waiting time for passengers.Mathematical model of double objective functionThis thesis designs a solution for bus scheduling optimization based on an improved genetic algorithm.In the process of optimizing the scheduling results,this paper improves and optimizes the algorithm in three stages: selection,crossover,and mutation.The improvement of selection is to design a dynamic fitness function and adopt the method of selecting multiple copies of excellent individuals without replacement.The improvement of the crossover is the design of a new crossover operator.The crossover operator takes into account the large difference between the quality of the initial group and the later group,so an adaptive crossover function is used.The improvement on mutation is the introduction of a tabu search algorithm.Based on the forecast of passenger flow,based on an improved genetic algorithm,the model is solved to determine the optimal schedule and minimum number of vehicles allocated.Through experimental example analysis,based on big data mining analysis,the probability of getting off the bus and the number of getting off the bus can be obtained.The regression analysis method can significantly predict the passenger flow.The improved genetic algorithm can well improve the bus scheduling problem.