Verification and Optimization of Metro Fare Clearing Models Based on Travel Route Reconstruction

How to verify and optimize metro fare clearing models efficiently and accurately is a research focus in metro operations. Metro fare clearing models are mostly based on probability distributions. In such models, the normal distribution of travel time corresponding to the section probabilities is used to calculate the route choice probabilities of passengers on a multiroute metro network. By integrating the operating mileage proportions of each metro line operator and the corresponding route choice probabilities, the fare clearing proportions are calculated for all the operators of the metro network. To verify the accuracy of the fare clearing proportions, we propose a travel route reconstruction approach based on cell phone data acquisition technique. With wireless access point (AP) sensors installed at transfer stations, the unique medium access control (MAC) address of the smart phone with Wi-Fi function turned on is recorded and transmitted to a data analysis platform. After matching the MAC address information with time and location, the travel route of the smart phone user is reconstructed. Then, the parameters in the fare clearing model are verified and optimized according to the travel route choice probabilities. The proposed methodology is applied in Hangzhou metro network for experiment, and the metro fare clearing model is verified and modified by reconstructing the actual travel routs of the local passengers. ARTICLE HISTORY Received: 16-07-2020 Revised: 10-09-2020 Accepted: 18-09-2020

The paper is arranged as follows. In Section 2, the recent studies conducted in both metro fare clearing and cell phone data based travel route reconstruction fields are reviewed. In Section 3, the metro fare clearing models in metropolitans of China are briefly introduced. In Section 4, a new metro travel route reconstruction approach is proposed. In Section 5, Hangzhou metro network is used as an example to verify and optimize the fare clearing model based on cell phone data acquisition. Finally, the conclusions are summarized in Section 6.

Literature review
The core algorithm in metro fare clearing model is the travel route reconstruction approach of passengers.
The reconstruction is based on building the connection between the travel route choice and its influence factors. In this section, both fare clearing model and cell phone based travel route reconstruction are discussed.

Metro fare clearing models
For metro fare clearing models, a consistent conclusion has been achieved (Zhao et al., 2007;Wang et al., 2013;Lu, 2012;Yu & Wang, 2013)  However, the passenger travel behavior survey in that paper is simply based on questionnaires, and the reliability of the survey results needs to be further verified.
In Zhou (2014), the actual route choices were used to revise the parameters in Shanghai metro fare clearing models. Since the metro network was really large and Yichao (2020) New Metro (2020) 1(1): 34-47 36 there are several routes in an origin-destination (OD) pair, it was not easy to figure out the actual OD routes.
The author used the entry-exit time of each passenger recorded by AFC system to match the most possible travel route, and regarded the obtained route as the actual OD one. However, this approach does not work for routes with similar travel time.
In Sun et al. (2015), an integrated Bayesian statistical inference framework was proposed to develop a passenger flow assignment model in a complex metro network. The integrated approach was applied to the metro network in Singapore, and the estimation of

Verification and optimization of fare clearing models
To illustrate the proposed methodology, we use Hangzhou metro network as an experimental site.

Background introduction
Till now, Hangzhou has three metro lines in operation.   According to the layout of the six stations, we decide the numbers of data acquisition devices as those listed in Table 2.

Acquisition and analysis of cell phone data
We conduct this experiment for 23 days from January Yichao (2020) 43 rational that the sampling rates at transfer stations are lower. In addition, one passenger may have more than one smart phones but only has one record in the AFC system, while his smart phone data will be captured more than once. Besides, the smart phone data of the working staff are not recorded in the AFC system but are recorded in the cell phone database. Therefore, it is rational that the sampling rate of Jiangling Road is over 100%.
After matching the acquired unique MAC address of each smart phone according to the record time and location, we can reconstruct the travel route and also calculate the sampling rates of patronage on the four typical OD pairs with multiple routes, as shown in figure 6.  The average sampling rate of patronage is 23%. As the AP sensors are forbidden to be installed in the track area, only a small amount of cell phone data can be recorded for the passengers on the train. To analyze the effects of different sampling rates on the route choice probability, a sensitivity analysis is conducted.
The detailed results are shown in Table 3.
From Table 4, one can figure out that the difference between the maximum and minimum sampling rates does not significantly affect the route choice probability. The travel route reconstruction system based on cell phone data is quite stable.

Verification of the fare clearing model
Referring to the reconstructed travel routes of the 23% Yichao (2020) 45 We conclude from Table 4  We use the least square estimation method to calibrate the empirical parameters in the Hangzhou metro fare clearing model on MATLAB. In Table 5, and are the transfer amplification coefficients for the 1st and 2nd transfer time, respectively.   licence, which permits copy and redistribute the material in any medium or format for any purpose, even commercially The licensor cannot revoke these freedoms as long as you follow the licence terms Under the following terms you must give appropriate credit, provide a link to the license, and indicate if changes were made You may do so in any reasonable manner, but not in any way that suggests the licensor endorsed you or your use If you remix, transform, or build upon the material, you may not distribute the modified material To view a copy of this license, visit https://creativecommons org/licenses/by-nd/4 0/