First Train Timetabling for Urban Rail Transit Networks with Maximum Passenger Transfer Satisfaction

Round 1

Reviewer 1 Report

The revision addresses my earlier concern about the survey approach. The added waiting time satisfaction function is on-point, and it offers scientific justification. Although I don’t think K-means is the best way to cluster the maximum tolerable waiting time, it serves the purposes here. 

Overall, I am satisfied with the revisions. This paper presents a mixed approach using both qualitative and quantitative methods that can potentially improve train timetable management in a practical way. The topic also suits this journal well. 

Author Response

Thanks for your comments. And also thanks for your previous suggestions that help us with the revisions on the manuscript.

Reviewer 2 Report

Dear authors,

I find your paper very interesting. I propose you some minor improvement in your paper:

1. When you used references in style for example line38 [2-10] its very unprofessional. Its mean that you add 9 references with no explanation why they crucial for your research. Similar line 56. Always is better to add one to two sentences about references that you use in your paper.
2. In abstract line 26 and 27 you stated “Finally, four practical suggestions are proposed 27 for urban rail transit network operations.” Please elaborate this in chapter 5 Case study.
3. Chapter 5 Case study must be concrete results of your methodology.
4. In table always algin numbers to left. In this why we can compare the results.
5. Figures are in bad resolution. For sure this is a technical issue but be aware of this issue.
6. Please add appendix after the references.

Regards,

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The publication focuses on the analysis of passenger preferences in the urban railway transport network. The authors present a new model that optimizes the schedule according to the preferences of waiting for a transfer to the next train line.

The first part is a sufficient introduction to the issue and presentation of related works. The following section describes the questionnaire survey and its results.
- How were the average values ​​calculated? Arithmetic or weighted average?
There is still a lack of necessary information proving the validity of the questionnaire survey and the ability to generalize the results.
- How did you measure the representativeness of the results of the questionnaire survey to allow generalization?
- Is there a 1: 1 passenger ratio on the given connections? And so on.

The chosen topic is very interesting and necessary to solve. It is beneficial to use your approach when designing/optimizing existing transport connections. Without a conclusive questionnaire survey that can be statistically verified and whose results can be generalized, your article only describes an interesting relationship of theories, which is insufficiently shown in the Case Study. Case Study is supported by a bachelor's degree questionnaire survey, without the possibility to generalize the results and make recommendations.

I understand that there is a lot of work in the background, but consider changing the goal of the publication.
1) Either present an original combination of theories and approaches - then focus on methods and reduce the importance of the results of the questionnaire survey or
2) you want to optimize transport connections - but then you have to carry out a significantly larger questionnaire survey.

Unfortunately, this text is an inappropriate combination of both, where interesting theories are insufficiently presented at the expense of an inconclusive questionnaire survey.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Thanks to the authors for answering my questions and doubts. The revised text now has a more stable basis, which is necessary for publishing the approach. The article is now suitable for publishing.

Author Response

Thanks for your comments. And also thanks for your previous suggestions that help us with the revisions on the manuscript.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The publication focuses on the analysis of passenger preferences of the urban rail transit network. The authors present a new model optimizing the timetable according to the preferences of waiting for the transfer.

 

The first part is a sufficient introduction to the issue and the presentation of related works. The following section describes the questionnaire survey and its results. However, there is a lack of information on data representativeness, questionnaire returns, population size, data distribution, and data validity. The authors presented the results of the Cluster analysis. Figure 4 and Table 2 show the same data.

Results showing that for 90% of passengers a 0 minute waiting for a connection is stressful is an unsurprising and expected finding.

 

The questionnaire provides outputs of a psychological rather than quantitative nature. The stated values ​​of the maximum tolerant waiting time and the most comfortable waiting time are instead a by-product and without further analysis do not provide a sufficient basis for the model. Many aspects have not been taken into account - daily time, the purpose of driving, passenger health, etc.

The created model draws on some unverified data, which are assigned to influence the objective function. The model is then optimized using the Bee colony algorithm.

 

The case study shows the number of passengers without giving details of the values ​​obtained.

The chosen topic is very interesting and necessary to solve, but unfortunately, the authors chose an unsuitable basis. The following results are just a transmitted error and presentation of the results that are expected after optimization.

 

I recommend the authors to make a more in-depth initial questionnaire survey and to analyze optimization approaches, such as Fuzzy.

Reviewer 2 Report

Dear Authors,

Overall your paper is interesting but here are some questions need to be raised and propose some improvement of paper:

1. How you made a questionnaire (on spot, internet, …)
2. You have overall 251 valid questionnaires. Is this enough? How big is sample size?
3. Always first explain picture/figure/table and then show them.
4. Figure 1 is surplus. One sentence is enough.
5. Line 145 – dimension of 5.42?
6. Check Figure 3 – when a have add I cannot get 1!
7. Explain differences between Figure 3 and Figure 4.
8. In tables please align all numbers to right!
9. Figure 9 is difficult to read.
10. Please make new chapter (or subchapter) discussion about your results.
11. Please give a straightforward goal of your paper in sustanibility!

Regards,

Reviewer 3 Report

The paper presented a novel solution to optimize transit timetable in terms of waiting time satisfaction. I found the paper is well organized, the methodological framework is clear, and the results look promising. However, I still have a few critiques that may help the authors with the revision.

First, a potentially major drawback of the method is in the design and conduct of the survey. The survey only questioned people about their tolerance of waiting time, regardless of the station type and other conditions: major station or small station; transfer station or not; the frequency of trains; rush hour or not; purpose of travel. The waiting satisfaction strongly associated with the context of when and where people are waiting for the next train. For example, people in the morning rush hour may be more anxious about missing the train especially if the frequency is relatively low. So the design of the questionnaire has flaws in my opinion. The conduct of the survey, i.e. the selection of people to take the survey, can be improved as well. A better way to conduct the survey is to directly ask people who are waiting for the train at various types of stations at different times of a day. Since the survey results serve as the foundation of the latter model building and data analysis, this could be a major drawback.

Second, I don’t think K-means is the best way to cluster the maximum tolerable waiting time. In Figure 4, the cluster near 20 mins is obviously too close to the cluster near 25 mins. There is no clear difference between these two clusters. K-means also requires the preset number of clusters as an input parameter, which doesn’t seem appropriate here. The purpose of cluster analysis is simple here, which is to classify the tolerable time from the survey results. I believe there are better ways to do that, like natural breaks, density-based clustering, or even classification based on standard deviation.

Third, page 6 line 180 “We believe that this feature coincides with people’s psychological changes with respect to time satisfaction.” It lacks scientific justifications here. Are there theories from behavioural psychology that can help back your choice?

Fourth, the paper can be shortened by putting some of the content into the appendix.