Research Of Urban Rail Transit Scheduling Optimization Based On The Demanding Of Passenger Flow

With China’s rail transit network in major cities being formed and perfected, it’s gradually appeared that the rail transit operation management has gained more importance and will face more requirements and challenges. In view of the current situation that the optimization and adjustment of train schedules are significantly subjected to the forecast accuracy of section passenger flow while the historical operating data fails to be rally used, based on the historical transaction records kept by the urban rail transit AFC (Automatic Fare Collection) system, the extraction of traffic demands of a single railway under network environments is explored in this paper, and the constructing and solving method of the scheduling optimization model are also studied.Above all, the current research on transportation scheduling optimizations, both at home and abroad, is generally reviewed in this paper. Then, the AFC system and the acquisition of trading information is briefly introduced while the extraction schemes and implementation steps for uplink traffic information of metro line 2 are proposed on the case of nanjing subway network; Especially, the "generalized OD" and "associated OD" concepts are given, which provide a powerful tool for network traffic distribution. Based on the set of corrected sample of the associated OD of metro line 2, the extraction solutions for uplink platform passenger flow is proposed, and the MATLAB implementation are also given for the typical stations. Meanwhile, on the basis of traditional OD matrix, a concept of "one-way traffic probability transfer matrix" is put forward while data mining technology is also used for the clustering analysis of uplink traffic probability transfer matrix, inorder to find a reasonable division to the operation day and operation periods. Finally, based on the divided operation periods and pre-extraction of traffic demands, train scheduling optimization model is built while time-sharing departure intervals are selected as the decision variables; With wide applicability and strong robustness, the Pseudo-Parallel Genetic Algorithm, which can effectively overcome the premature phenomenon, is selected to solve the optimization model, and the scheduling index under the optimized timetable is compared with the equivalent of the actual schemes to verify the effectiveness of the optimization.In general, based on the historical transaction information collected by the AFC system, the scheduling optimization scheme for a single railway under network environment is given in this paper and some beneficial exploration has also been made in data mining and usage of historical trading information, providing new thoughts of daily adjustment for schedules.