Research On Urban Bus Dispatching Based On Robust Optimization

In recent years,China’s urban transportation problems have increasingly become prominent,and vigorously developing public transportation is still considered the main method to solve urban transportation problems.However,in the current research of bus dispatching,most of the future passenger flows are simulated and predicted as much as possible,and then a model is established to solve the problem based on the accurate passenger flow data obtained.Even if the real-time passenger flow data obtained is completely accurate,there are still uncertain changes in passenger arrival rates.Robust optimization is a commonly used and effective method to deal with uncertainty problems.This dissertation introduces a robust optimization method to the field of public transportation dispatching to study the urban public transportation vehicle dispatching plan under the uncertain passenger arrival rate.It improves the bad conditions in the public transportation operation process and provides high-level public transportation for urban resident service.The research of this article mainly includes the following three aspects:1.Establish a robust optimization model for the interval-based single-line bus departure interval.The uncertainty of passenger arrival rate at bus stops is described as the number of intervals,allowing the passenger arrival rate to fluctuate within a certain range.The intervalbased robust optimization method is adopted,and the conservative degree of the parameter adjustment model is introduced.Taking the weighted sum of the average passenger waiting time and the average passenger volume as the objective function,considering the logical constraints and intermediate process constraints in the actual bus operation process,a robust optimization mathematical model of the interval-based single-line bus departure interval is established.2.Design an improved genetic-simulated annealing algorithm to solve the model.Combining the characteristics of the model,the overall idea of the genetic-simulated annealing algorithm and the various sub-modules which include coding method,genetic operation,and simulated annealing cooling operation are improved.The improvements include designing coding schemes,combining deterministic selection operations and the best retention mechanism through different parents constructing the progeny population.Adaptive probability and improved cooling function are adopted to form an improved GA-SA algorithm to improve the operating efficiency and optimization performance of the algorithm.3.Combine examples to verify the applicability of the method.Select a certain line in Zhengzhou as a case object,using MATLAB to solve the established optimization model.Design experiments to verify the applicability of the robust optimization results when the actual passenger arrival rate meets the common distribution.The results show that the robust solution can effectively control operational risks of bus companies,reducing the average waiting time for passengers,and improve the quality of travel services.In addition,the model can be used as a sub-model to be nested into the bus network flow distribution model and bus route design model full of application potential.