Research on Dynamic Subsidy Based on Deep Reinforcement Learning for Non-Stationary Stochastic Demand in Ride-Hailing

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article has the right structure.

The cited literature items are relevant and up-to-date.

The model used seems to be well implemented for the purposes of the article.

The purpose of the study is properly presented and, in my opinion, was properly implemented.

The research results are properly presented and the conclusions are correctly formulated.

The only recommendation is to add a "Discussion" section. However, here I leave the choice to the authors. I ask the authors to consider whether adding a discussion before the "Conclusion" section would add value to the article by citing similar studies and comparing them.

The language of the article is correct. No major stylistic errors.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

General comments: The manuscript presents a study on dynamic subsidies in the taxi market using deep reinforcement learning to address the problem of unstable stochastic demand. The work is well structured and the methodology is scientifically sound, providing valuable insights for optimizing subsidies to increase platform profitability and mitigate the imbalance between supply and demand.

Specific comments:
- Correct line spacing. Unify the line spacing in sections lines 166-178, 256-298, 338-371, 402-405, and others to match the rest of the document.
- Increase the font size in the figures. To improve readability and ease of interpretation, increase the font size in all figures, tables, and diagrams.
- Correct the reference to the data source. Provide a correct reference to the data source mentioned in the data availability statement so that the data presented can be properly evaluated and verified.
- Further research. The conclusions section does not provide directions for further research. It would be useful to indicate possible directions for future research to provide readers with an idea of the prospects and further development of the topic.
- Application of the algorithm. It would be interesting to add a discussion on the possible use of the proposed algorithm in other contexts or industries where there is also volatile demand and the need for dynamic optimization.

Overall, this is an interesting and well-organized paper. The proposed dynamic subsidy strategy is scientifically sound and has practical potential in the taxi market. Correction of specific comments on line spacing and incorrect reference to the data source would further improve the manuscript.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

 

Thank you for the opportunity to review this article. The manuscript addresses a topic of great relevance, given the growing dependence on private transport services and the need for efficient demand management. The literature review is comprehensive, covering several pertinent studies and positioning current research in the context of existing work, highlighting gaps that this study aims to fill. The methodology details the construction of the NSMDP model and the implementation of the deep reinforcement learning algorithm. The research has significant practical implications for private transport platforms, potentially improving service profitability and efficiency through more effective subsidy strategies.

 

Some suggestions to improve the manuscript:

 

1. Some sections contain grammatical errors and sentence constructions that may make understanding difficult. A thorough review and editing process is recommended to improve the clarity of the text.

2. The abstract is extensive and includes detailed methodological information that may overwhelm readers. It is suggested to condense the summary to focus on the main objectives, methods, results and implications of the study, seeking greater brevity and impact.

3. The introduction provides good context but could be more engaging to capture the reader's interest from the beginning. It is important to clearly present the research problem and its relevance at the beginning.

4. Although the methodology is detailed, it could benefit from visual aids, such as flowcharts or diagrams, to illustrate complex processes. Including visual representations of the NSMDP model, the deep reinforcement learning framework, and the change point detection method can improve understanding.

5. The results section is comprehensive, but lacks a critical discussion of the implications, limitations, and potential biases of the findings.

6. The literature review table is useful but could be more informative. It is suggested to include a summary of the main conclusions or contributions of each referenced study to provide more context.

7. The conclusion summarizes the study well but may offer further directions for future research.

By addressing these points, the manuscript can be significantly improved, increasing its chances of acceptance.

Comments on the Quality of English Language

Some sections contain grammatical errors and sentence constructions that may make understanding difficult. A thorough review and editing process is recommended to improve the clarity of the text.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors responded to the suggestions satisfactorily and improved the quality of the article. The manuscript may be accepted.

Comments on the Quality of English Language

The authors made corrections to the text and quality of english language improved