Bus Dynamic Scheduling Research For Transfering Hub

Transfer hub site for the role of public transport network is similar to the artery for the role of the person's life. Its good association directly affect that the public transport system service level whether good or bad. Only efficient,convenient and comfortable service helps to improve the proportion of bus in the traffic transportation, makes the traveler will take the bus as the first choice of travel vehicles. The research emphasis in this paper is that how to optimize the bus dynamic scheduling in the transfer station based on the bus arrival time. In order to reach its good convergence ability.First, the paper analyzes the different forms of public transport scheduling. According to the characteristics of bus scheduling in transfering hub site, it is given the necessity of implementing dynamic scheduling in transfering hub site. It introduces the classification of the bus schedule based on the different standards. Between bus routes different connecting ways lead to different ways of transfer, and analyses the two different transfer mode, so as to passenger in the transfer station use the appropriate tramsfer mode in this paper(transfer on the same bus station).Second, because the public transport vehicles are susceptible to the influence of traffic emergency on the road, the public transportation scheduling problem turns into a very complicated problem. By bus scheduling strategy can minimize this type of interference to the impact on the running bus, so that can reach the bus back to the original normal operating environment. Based on bus real-time information directly decided the characteristics of the effectiveness of the scheduling policy is proposed based on dynamic bayesian- least squares support vector machine(SVM) bus arrival time prediction model. And updated data in time by using the method of increase or decrease the samples. Which enables the input data to better and more accurately reflects the current bus situation. And the arrival time of the bus 15 road in lanzhou can be predicted dynamicly.Comparing with the traditional LS-SVM, the analysis shows that, based on dynamic Bayes theorem and LS-SVM arrival time prediction method can get the result of the high precision relatively.Finally,tansfer between bus lines through the transfer hub accurately, rapidly, coordinatedly. which is very important for the transfer hub to improve the function and quality of service. So as to maximize the efficiency of transportation in return for achieving the lowest investment. In this article the bus dynamic scheduling model is established under the circumstances which the bus arrival time and departure time are known(via Bayes theorem and LS-SVM prediction analysis). In order to obtain minimum the cost sum of the travelers' waiting,transfer and bus operation. The model constraints include the bus arrival time, relaxation time of bus in the initial site, the bus service frequency to ensure the coordination between the lines scheduling. In view of the specialty of the model and algorithm is given based on the NSGA-? to solve the model. Finally, the dynamic scheduling model of 5 lines in Lanzhou are the hub site transfer relationship is given. The results of analysis showed that in the implementation of dynamic scheduling according to the original schedule after the operation of the line to save 16.9%.