Dyn-Pri: A Dynamic Privacy Sensitivity Assessment Framework for V2G Interactive Service Scenarios

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The work is interesting and addresses a topic of current relevance. However, I found the structure used by the authors to describe the study somewhat confusing. Another point of concern is that it is not clear how the authors arrived at the results presented. It might be helpful to provide a more detailed description of the methodology used. Finally, I believe the conclusion is too brief, lacking a discussion of the current study’s limitations and suggestions for future work.
Below are some more specific recommendations.

Figures 2  and 6 – The citation of Figure 2 and Figure 6 in the text only appears after it is presented. I suggest that the authors refer to the figures before its presentation. The same regarding figure 8.

The authors refer to the different figures in the manuscript sometimes as “Fig” and sometimes as “Figure.” I suggest standardizing the terminology.

What does CDF  and POI means?

 

 

Author Response

Comments 1: The work is interesting and addresses a topic of current relevance. However, I found the structure used by the authors to describe the study somewhat confusing. Another point of concern is that it is not clear how the authors arrived at the results presented. It might be helpful to provide a more detailed description of the methodology used. Finally, I believe the conclusion is too brief, lacking a discussion of the current study’s limitations and suggestions for future work.

Response 1: 

We sincerely thank the reviewer for the positive feedback and for the constructive comments.

In response to the concern about the manuscript’s organization, we have now explicitly clarified its structure and core contents in the revised version. To demonstrate the effectiveness of our proposed Dyn-Pri method, we provide comprehensive experimental validations involving four distinct EV-user profiles—characterized by varying privacy preferences and awareness levels—across four representative V2G services within three typical V2G scenarios. These experiments not only validate the feasibility of Dyn-Pri but also illustrate its advantages for supporting V2G operations, particularly in demand-side response (DR). Furthermore, we have expanded the concluding section to include a thorough discussion of the study’s limitations and to outline directions for future research.

Comments 2: Figures 2 and 6 – The citation of Figure 2 and Figure 6 in the text only appears after it is presented. I suggest that the authors refer to the figures before its presentation. The same regarding figure 8.

Response 2: We agree with your suggestion. We will adjust the placement of figure citations by adding brief references to Figure 2 (e.g., “As shown in Figure 2, the geographical distribution of charging stations in Shenzhen forms the basis of our experimental dataset”) just before its presentation in Section 3.1. Similarly, we will insert pre-presentation mentions for Figure 6 and Figure 8 in Section 3.4 (e.g., “Figure 6 illustrates the privacy sensitivity differences under varying charging station adjacencies”) to align with the suggested citation order, which involves only minor text adjustments.

Comments 3: The authors refer to the different figures in the manuscript sometimes as “Fig” and sometimes as “Figure.” I suggest standardizing the terminology.

Response 3: Thank you for pointing out this inconsistency. We will uniformly revise all figure references to “Figure” throughout the manuscript, which is a straightforward editing task that does not affect the original content.

Comments 4: What does CDF and POI means?

Response 4: We apologize for the oversight. We will add parenthetical definitions for CDF (Cumulative Distribution Function) and POI (Point of Interest) at their first occurrences in the text: CDF will be defined in Section 3.3 when first mentioned (e.g., “Cumulative Distribution Function (CDF) of privacy sensitivity”), and POI will be clarified in Section 3.4 to ensure clarity without expanding the text unduly.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors of this quite interesting paper propose Dyn-Pri, a multidimensional privacy sensitivity assessment framework for V2G interactive service scenarios. Dyn-Pri addresses the limitations of existing privacy sensitivity assessment methods by incorporating a comprehensive multi-dimensional quantification model that integrates objective and subjective privacy elements. In addition, it considers the dynamic coupling relationships among these elements. 

The subject of the paper is interesting and very well presented. The analysis is correct, the results presented are adequate and the same applies to the related work. The paper can be accepted for publication but the authors should correct some syntax/grammar errors that appear in the paper. 

Author Response

Summary: The authors of this quite interesting paper propose Dyn-Pri, a multidimensional privacy sensitivity assessment framework for V2G interactive service scenarios. Dyn-Pri addresses the limitations of existing privacy sensitivity assessment methods by incorporating a comprehensive multi-dimensional quantification model that integrates objective and subjective privacy elements. The subject of the paper is interesting and very well presented. The analysis is correct, the results presented are adequate and the same applies to the related work. In addition, it considers the dynamic coupling relationships among these elements.

Response: Thank you very much for your positive evaluation and encouraging comments. We appreciate your recognition that the topic is “interesting and very well presented,” and that the analysis, results, and related work are “correct” and “adequate.” Your feedback confirms that Dyn-Pri effectively addresses the limitations of existing privacy-sensitivity assessment methods by integrating both objective and subjective dimensions in a multidimensional, scenario-aware manner.

Comments 1: The paper can be accepted for publication but the authors should correct some syntax/grammar errors that appear in the paper.

Response 1: We sincerely appreciate your positive evaluation and valuable suggestion. We will carefully review the entire manuscript, with a focus on checking for syntax and grammar errors. Special attention will be paid to the consistency of verb tenses in the methodology description and the accuracy of sentence structures in the experimental results analysis. This revision will ensure the accuracy and fluency of the expression.

 

 

Reviewer 3 Report

Comments and Suggestions for Authors

The paper examines the privacy sensitivity of the participants for Vehicle-to-Grid (V2G) services. The interconnected influences of subjective and objective privacy elements of information sharing during V2G service operations are analyzed by the authors.

The paper manages to address a crucial trade-off: efficient use of data without compromising privacy, which is vital for modern energy systems based on digital cooperation.

Integrating objective and subjective privacy elements, as well as the dynamic interactions between them, adds depth to the assessment of privacy risks.

Authors can improve the presentation of their manuscript by considering the following observations.

Explain abbreviations when they first appear in the text, even in the abstract. Avoid using abbreviations in the title if possible.

Repeating the phrase "Dyn-Pri" in almost every paragraph can be quite annoying for readers.

The study has some limitations because it does not clearly mention the computational effort required to implement the model in a real and dynamic large-scale V2G system.

Author Response

Summary: The paper examines the privacy sensitivity of the participants for Vehicle-to-Grid (V2G) services. The interconnected influences of subjective and objective privacy elements of information sharing during V2G service operations are analyzed by the authors. The paper manages to address a crucial trade-off: efficient use of data without compromising privacy, which is vital for modern energy systems based on digital cooperation. Integrating objective and subjective privacy elements, as well as the dynamic interactions between them, adds depth to the assessment of privacy risks. Authors can improve the presentation of their manuscript by considering the following observations.

Response: Thank you very much for your positive appraisal of our work. We sincerely appreciate your recognition of the importance of investigating privacy sensitivity among participants in V2G scenarios, as well as your commendation of the in-depth analysis we provided through Dyn-Pri. Your acknowledgment that we successfully integrate objective and subjective privacy elements—while capturing their dynamic interdependencies—reinforces our confidence in the depth and rigor of the framework. Moreover, we are grateful for your highlighting the critical balance between data utility and privacy protection, which is indeed central to the viability of modern, digitally-cooperative energy systems. Your encouraging remarks motivate us to refine the presentation further and address the constructive suggestions you have provided.

In addition, we also extend our sincere thanks for the three constructive suggestions you provided. We have carefully addressed each point and incorporated the corresponding revisions into the updated manuscript.

Comments 1: Explain abbreviations when they first appear in the text, even in the abstract. Avoid using abbreviations in the title if possible.

Response 1: We appreciate this suggestion. We will revise the text to explain all abbreviations when they first appear, including in the abstract. For example, "V2G" will be clarified as "Vehicle-to-Grid" at its first occurrence in the abstract and main text. Regarding the title, since "Dyn-Pri" is the core framework proposed in this study, we hope to retain it for clarity and recognition, but we will ensure all other abbreviations are avoided in the title.

Comments 2: Repeating the phrase "Dyn-Pri" in almost every paragraph can be quite annoying for readers.

Response 2: Thank you for pointing this out. We will carefully review the manuscript and reduce the frequency of "Dyn-Pri" repetitions. In appropriate contexts, we will use alternatives such as "the proposed framework" or "this assessment model" to enhance readability while ensuring the clarity of reference to our proposed framework.

Comments 3: The study has some limitations because it does not clearly mention the computational effort required to implement the model in a real and dynamic large-scale V2G system.

Response 3: Thank you for your valuable observation regarding computational effort in large-scale V2G systems. While the current manuscript does not explicitly elaborate on this aspect, the reinforcement learning-based dynamic optimization mechanism (detailed in Section 2.3) is inherently designed to adapt to large-scale scenarios by optimizing parameter update efficiency. Moreover, the experimental validations using a six-month dataset with 17,532 charging piles (Section 3.1) implicitly demonstrate the model's feasibility in handling large-scale dynamic data, as the framework successfully processed and analyzed massive interaction data under varying V2G service conditions.

Reviewer 4 Report

Comments and Suggestions for Authors

I appreciate the opportunity to review this manuscript. The research is relevant as it focuses on improving vehicle comfort and stability, which is a key factor in advancing innovation within the field of electromobility. Below, I outline a number of suggestions for the authors to consider.

Suggestions:

+ Introduction

• I recommend moving the paragraph beginning with “Existing studies mainly quantify… discharging control instructions.” (Lines 50–61) to Section 2, “Dyn-Pri: a privacy sensitivity assessment framework for V2G interaction service scenarios”. This would allow for a more coherent integration with the literature review and conceptual framework, improving the flow and organisation of the manuscript.
• It would also be beneficial to expand the introduction by providing more context about the actors or elements involved in the system under study. This would offer a clearer understanding of the current situation in the field and help establish a stronger connection with the research approach.

+ Discussion Section

I suggest including a dedicated discussion section, or alternatively, combining the results and discussion into a single section. This would help to critically interpret the findings and reinforce the central arguments. At present, Section 3, “Experimental Validations and Analysis”, contains too much information without clearly distinguishing between the results and their analysis, which weakens the main line of argument.

+ Conclusions

While the conclusions are consistent with the study’s objective, they could be strengthened by engaging more critically with the literature reviewed. Doing so would reinforce the claims derived from the simulations and provide greater robustness to the overall findings.

Formatting

• To improve the readability of the abstract, it is suggested to avoid repeating the transition word “However” in both line 10 (“However, the data quality…”) and line 11 (“However, there is…”). Consider using alternative expressions.
• I recommend that the authors check the formatting of all subheadings to ensure consistency with the journal’s guidelines. For example, in lines 107 and 399, there are inconsistencies, as some subheadings appear in bold while others do not.

Author Response

Summary: I appreciate the opportunity to review this manuscript. The research is relevant as it focuses on improving vehicle comfort and stability, which is a key factor in advancing innovation within the field of electromobility. Below, I outline a number of suggestions for the authors to consider.

Response: Thank you very much for reviewing our manuscript and for recognizing the relevance of our research to vehicle comfort and stability—key drivers of innovation in electromobility. We appreciate your constructive suggestions and have carefully revised the paper to address each of them. Your feedback has helped us improve both the clarity and rigor of our work.

 Comments 1:

Introduction
• I recommend moving the paragraph beginning with “Existing studies mainly quantify… discharging control instructions.” (Lines 50–61) to Section 2, “Dyn-Pri: a privacy sensitivity assessment framework for V2G interaction service scenarios”. This would allow for a more coherent integration with the literature review and conceptual framework, improving the flow and organisation of the manuscript.
• It would also be beneficial to expand the introduction by providing more context about the actors or elements involved in the system under study. This would offer a clearer understanding of the current situation in the field and help establish a stronger connection with the research approach.

Response 1: Thank you for your insightful suggestion. After careful consideration, we have decided to retain the original overall structure of the paper; nevertheless, we fully appreciate the reviewer’s underlying intent in suggesting that the paragraph beginning with “Existing studies mainly quantify…” (Lines 50–61) be moved into Section 2. To foreground the need for Dyn-Pri and the specific problems it tackles, we have therefore added a dedicated “Motivation and Problem Statement” subsection at the outset of Section 2, which incorporates and expands upon the content of that paragraph. Regarding the introduction section, we have supplemented the innovation of existing methods to clarify the system context and strengthen the connection with our research approach.

Our revisions are as follows:

“2.Dyn-Pri: a privacy sensitivity assessment framework for V2G interaction service scenarios”, p.3 (Lines 106–116)in the revised paper.

Privacy sensitivity is the degree of risk or concern from potential leakage of shared data, shaped by both objective data attributes and subjective factors like user perceptions and contextual dynamics. Existing literature predominantly quantifies privacy sensitivity with information-entropy, differential-privacy, or set-pair-analysis models. While these techniques yield useful benchmarks, they are inherently static since they treat data attributes such as charging power or State-of-Charge as fixed inputs and neglect the dynamic, multi-agent realities of V2G ecosystems. Specifically, they overlook time-varying contextual factors and the strategic interactions among heterogeneous stakeholders. Consequently, current metrics cannot adapt to the continuously evolving, collaborative requirements of V2G services. This gap motivates us to design a new framework that explicitly models both the dynamic context and multi-agent privacy game.

Comments 2:

Discussion Section
I suggest including a dedicated discussion section, or alternatively, combining the results and discussion into a single section. This would help to critically interpret the findings and reinforce the central arguments. At present, Section 3, “Experimental Validations and Analysis”, contains too much information without clearly distinguishing between the results and their analysis, which weakens the main line of argument.

Response 2: Thank you for your valuable suggestion regarding the discussion section. In Section 3, “Experimental Validations and Analysis,” we conduct targeted experiments that cover four EV-user privacy profiles across four distinct V2G services within three representative scenarios. These validations are already organized into purpose-specific subsections that both present the results and interpret their implications in context. In light of this structure, we believe a separate Discussion section is unnecessary.

Comments 3:

Conclusions
While the conclusions are consistent with the study’s objective, they could be strengthened by engaging more critically with the literature reviewed. Doing so would reinforce the claims derived from the simulations and provide greater robustness to the overall findings.

Response 3: We agree with your suggestion. We will revise the conclusion to explicitly contrast Dyn-Pri with existing methods, emphasizing how our framework addresses their limitations. This critical engagement will reinforce the novelty and robustness of our findings, linking simulations to gaps identified in prior research.

Comments 4:
Formatting
• To improve the readability of the abstract, it is suggested to avoid repeating the transition word “However” in both line 10 (“However, the data quality…”) and line 11 (“However, there is…”). Consider using alternative expressions.
• I recommend that the authors check the formatting of all subheadings to ensure consistency with the journal’s guidelines. For example, in lines 107 and 399, there are inconsistencies, as some subheadings appear in bold while others do not.

Response 4: We appreciate your attention to formatting. In the abstract, we will revise the second “However” to “Notably” to avoid repetition and enhance readability. Regarding subheadings, we will thoroughly check all to ensure consistent formatting in line with the journal’s guidelines.

 

Reviewer 5 Report

Comments and Suggestions for Authors

The concept of “privacy sensitivity” is well structured, but the initial formulations could benefit from further clarification for readers unfamiliar with the field.

The Dyn-Pri proposal is innovative by integrating objective and subjective dimensions, but its contribution to existing methods could be highlighted more clearly in the introduction.

The model of interdependence between privacy elements is well-constructed mathematically and justified in the V2G context.

Experiments using real data from Shenzhen support the feasibility of the model; however, a direct comparison with other existing methods would strengthen the conclusions.

The application in real V2G scenarios is valuable, but details on how Dyn-Pri can be implemented in existing systems are lacking.

The paper does not explicitly discuss the computational complexity of the Dyn-Pri method, an important aspect for its broad applicability.

The graphical presentation of the results is useful, but the interpretation of some of the graphs (e.g. Fig. 5-7) could be better summarized in the text.

Although the content is rich, the language can sometimes be redundant. Recommendations for more concise expression would improve readability.

The introduction of reinforcement learning into the model is interesting, but details about learning parameters and convergence are vague.

The paper focuses on V2G scenarios in China; a discussion of the model’s portability to other markets/regions would be useful.

The sections are well-organized, but some technical subsections (e.g. 2.1.2 and 2.1.3) are dense and could benefit from interim summaries.

The limitations of the proposed method are not discussed; their inclusion would provide a more solid academic balance.

Author Response

Comments 1: The concept of “privacy sensitivity” is well structured, but the initial formulations could benefit from further clarification for readers unfamiliar with the field.

Response 1: We sincerely thank the reviewer for this suggestion. In the first paragraph of Section 2 of the revised manuscript, we have added a concise definition of privacy sensitivity, i.e., Privacy sensitivity is the degree of risk or concern from potential leakage of shared data, shaped by both objective data attributes and subjective factors like user perceptions and contextual dynamics. This definition is helpful for readers to immediately grasp the privacy concerns and impacts associated with data sharing among V2G participants.

Comments 2: The Dyn-Pri proposal is innovative by integrating objective and subjective dimensions, but its contribution to existing methods could be highlighted more clearly in the introduction.

Response 2: Thank you for this invaluable feedback. In Section 2 of the revised manuscript, we now explicitly position Dyn-Pri against the state of the art: whereas prior static, attribute-centric metrics such as information-entropy or differential-privacy models treat data features (e.g., SoC, charging power) as fixed and neglect the time-varying realities of V2G services, Dyn-Pri is the first framework to simultaneously integrate (1) objective data attributes, (2) operational context, and (3) user security experience, and to learn their real-time, bidirectional interdependencies via reinforcement learning. This contrast is now highlighted in a dedicated paragraph, clearly demonstrating how Dyn-Pri advances beyond existing methods.

Additionally in Section 2 of the revised manuscript, we have added a concise, high-level summary of Dyn-Pri’s innovations.

 Our revisions are as follows:

“1. Introduction”, p.2 (line 80-96) in the revised paper.

In summary, the main contributions of the paper are as follows.

• Three-Element Quantification Privacy Sensitivity We decompose privacy sensitivity into (I) objective data attributes, (C) context-driven subjective risk, and (P) participant-experience-based subjective risk, quantifying each with dedicated indicators. Its novelty lies in the unification of objective + dual subjective factors.
• Dynamic Intertwining Extension. We extend the static three-element model by explicitly capturing and learning the real-time, bidirectional influences among I, C, and P, replacing fixed weights with reinforcement-learning-driven adaptive mappings. Its novelty lies in the explicit modeling of time-varying bidirectional effects)
• Self-Optimizing Holistic Model. We integrate the intrinsic scores and learned intertwining effects into a single holistic score that self-optimizes across varying V2G scenarios, validated against four EV-user privacy profiles in four distinct V2G services across three operational scales, demonstrating superior accuracy and demand-response support. Its novelty lies in the closed-loop, scenario-aware calibration)

“2. Dyn-Pri: a privacy sensitivity assessment framework for V2G interaction service scenarios”, p.3 (line 108-116) in the revised paper.

…Existing literature predominantly quantifies privacy sensitivity with information-entropy, differential-privacy, or set-pair-analysis models. While these techniques yield useful benchmarks, they are inherently static since they treat data attributes such as charging power or State-of-Charge as fixed inputs and neglect the dynamic, multi-agent realities of V2G ecosystems. Specifically, they overlook time-varying contextual factors and the strategic interactions among heterogeneous stakeholders. Consequently, current metrics cannot adapt to the continuously evolving, collaborative requirements of V2G services. This gap motivates us to design a new framework that explicitly models both the dynamic context and multi-agent privacy game.

Comments 3: The model of interdependence between privacy elements is well-constructed mathematically and justified in the V2G context.

Response 3: We sincerely appreciate your positive appraisal of the interdependence model. In the revised manuscript we retain its exact mathematical formulation (Figure 1 and Equations 14–17) and now augment each equation with concise inline remarks that explicitly link the symbols to real-world V2G situations—for example, how an uptick in users’ privacy awareness instantly reshapes the data-sharing context—thereby rendering the justification both transparent and compelling.

 Comments 4: Experiments using real data from Shenzhen support the feasibility of the model; however, a direct comparison with other existing methods would strengthen the conclusions.

Response 4: Thank you for this invaluable feedback.

Using the UrbanEV dataset from Shenzhen to validate Dyn-Pri offers several distinct advantages. At first, the UrbanEV dataset rigorously collects 73 distinct data parameters in accordance with the China standard GB/T 32960.3-2016, providing the most comprehensive shared data foundation for all types of V2G services. Additionally, spanning an extensive six-month period, it covers a timeframe broad enough to capture seasonal and behavioral variations. Most important, Shenzhen—one of China’s megacities—hosts a massive EV user base across 1,362 charging stations equipped with 17,532 charging piles, making it an ideal and representative testbed for nationwide V2G initiatives.

Moreover, because existing alternative approaches are inherently static—they treat attributes such as charging power or State-of-Charge as fixed inputs and ignore the dynamic, multi-agent realities of V2G ecosystems—direct comparison with Dyn-Pri would place them at a clear disadvantage. For this reason, comparative results against those methods have been omitted.

 Comments 5: The application in real V2G scenarios is valuable, but details on how Dyn-Pri can be implemented in existing systems are lacking.

Response 5: Thank you for highlighting the need for implementation guidance. Implementation details for Dyn-Pri vary with the V2G service type and the number of participants. In a Residential-level standalone EV-charging scenario, owners rely solely on private, in-community piles; the corresponding services—billing summaries, load-scheduling optimization, and off-peak guidance—involve limited peer-to-peer data exchange and a small participant pool. Consequently, the (I) objective data attributes, (C) context-driven subjective risk, and (P) participant-experience-based subjective risk we must monitor differ markedly from those in a County-level single-operator network or a City-level multi-operator ecosystem, where larger scales and richer interactions heighten both dynamic complexity and contextual demands. Because these differences preclude a one-size-fits-all deployment roadmap, we have opted not to provide a case-specific implementation blueprint.

 Comments 6: The paper does not explicitly discuss the computational complexity of the Dyn-Pri method, an important aspect for its broad applicability.

Response 6: We acknowledge this oversight. Therefore, we have added a discussion of the Dyn-Pri’s computational complexity in Section 3.5.

  Our revisions are as follows:

“3.5. The Computational Complexity Analysis of Dyn-Pri”, p.15 (line 599-612) in the revised paper.

3.5. The Computational Complexity Analysis of Dyn-Pri

The computational complexity of Dyn-Pri mainly consists of three parts. Firstly, the basic quantification module involves weighted summation of multi-dimensional indicators when calculating the objective privacy sensitivity (I), context-related subjective sensitivity (C), and participant experience-related subjective sensitivity (P), with a time complexity of O(n), where n is the number of data attributes or participants. Secondly, the dynamic intertwining influence module captures the dynamic relationships among the three elements through reinforcement learning, whose core is the neural network approximation of the Q-value function, resulting in a time complexity of O(k·m), where k is the number of training iterations and m is the dimension of the state space. Thirdly, the overall assessment model integrates the results of the above two parts, with a time complexity of O (1). In summary, the total time complexity of Dyn-Pri can be expressed as O (n + k·m), where n is the scale of basic data, and k and m are related to reinforcement learning training.

Comments 7: The graphical presentation of the results is useful, but the interpretation of some of the graphs (e.g. Fig. 5-7) could be better summarized in the text.

Response 7: Thank you for this valuable suggestion.

For Figure 5, we in lines 500–506 of our manuscript explicitly summarize the key takeaway, i.e., this finding is crucial for participants seeking to balance data utility and privacy protection in V2G services, offering clear guidance on optimizing data-sharing practices without sacrificing operational efficiency.

For Figure 6, we have added a concise summary (lines 536–543) that interprets how varying spatial-adjacency granularity (quantified by hop count) influences Dyn-Pri's privacy-sensitivity assessments across V2G scenarios.

Figure 7 complements Figure 6 by demonstrating the impact of intertwining influences in a city-level Service-3 scenario. Its corresponding summary appears in lines 556–559, which can be considered as the  deepened validation initiated in Figure 6(b).

Figure 8 further consolidates the findings of Figures 6 and 7 by jointly examining the effects of intertwining influences and spatial adjacency; it serves as a confirmatory extension of the preceding analyses.

Collectively, the conclusions drawn from Figures 6–8 are not only restated within their individual captions and surrounding text, but also synthesized in lines 573–584 under Section 3.4's unifying theme“Validations of Privacy-Sensitivity Differentiation under Intertwining Influences.”.

   Our revisions are as follows:

“3.4 The validations of privacy sensitivity differentiation in the intertwining influences”, p.15 (line 543-550) in the revised paper.

In summary, Figure 6 demonstrates how spatial adjacency (hop count) modulates privacy sensitivity in county-level versus city-level V2G scenarios. The limited station interconnection under a single-operator county V2G service scenarios produces smaller hop counts and hence modest sensitivity differences while the richer spatial charging service links under the multi-operator city V2G service scenarios can provide markedly higher and more dispersed sensitivity scores as hop count increases. Obviously, Dyn-Pri successfully captures these regime-specific disparities.

Comments 8: Although the content is rich, the language can sometimes be redundant. Recommendations for more concise expression would improve readability.

Response 8: Thank you for this observation. We have performed a thorough language edit, eliminating redundant phrases, merging overlapping sentences, and replacing verbose constructions with concise equivalents. These changes substantially improve readability while fully preserving all technical content.

Comments 9: The introduction of reinforcement learning into the model is interesting, but details about learning parameters and convergence are vague.

Response 9: Thank you for highlighting this point.

  Our revisions are as follows:

“3.1 …. 4）The default hyper-parameters of reinforcement-learning during Dyn-Pri’s self-optimization”, p.10 (line 442–448) in the revised paper.

4）The default hyper-parameters of reinforcement-learning during Dyn-Pri’s self-optimization

Here, the default hyper-parameters of the reinforcement-learning are set as follows, i.e., learning rate 1×10⁻³, discount factor 0.95, replay-buffer size 50 000, and a double-DQN with two hidden layers (128×64 ReLU units). We also include the convergence criterion, i.e., training stops when the rolling-window average reward changes by less than 0.1 % over 5 000 episodes. This information ensures full reproducibility. 

 Comments 10: The paper focuses on V2G scenarios in China; a discussion of the model’s portability to other markets/regions would be useful.

Response 10: Thank you for this valuable suggestion. Although our validation experiments focus on Shenzhen, China, the Dyn-Pri framework is equally applicable to V2G ecosystems worldwide—from the EU to North America—where regulators and operators share the same dual mandate: safeguard user privacy in accordance with GDPR or equivalent statutes while ensuring that data-driven efficiencies are not sacrificed.

Comments 11: The sections are well-organized, but some technical subsections (e.g. 2.1.2 and 2.1.3) are dense and could benefit from interim summaries.

Response 11: Thank you for pointing out that Sections 2.1.2 and 2.1.3 are technically dense. To improve clarity, we have inserted concise interim summaries at the end of each subsection. These one-paragraph recaps restate the key take-aways in plain language and explicitly link the mathematical details back to the overall Dyn-Pri framework, ensuring the reader never loses sight of the main storyline while preserving all technical rigor.

  Our revisions are as follows:

“2.1.2. Subjective privacy sensitivity related to V2G service operational context”, p.6 (line 245–251) in the revised paper.

To summarize, this subsection establishes a quantification model for context-related subjective privacy sensitivity (C) by integrating three key indicators, i.e., privacy security trustworthiness (C_t) of the V2G operational context, data interdependence level (C_r) between participating entities, and data interdependence sensitivity (C_s). These indicators collectively capture the complex influence of multi-layered data-sharing relationships and service interdependencies among V2G participants on privacy sensitivity, with the final value of C derived through weighted summation of  C_t and C_s.

Comments 12: The limitations of the proposed method are not discussed; their inclusion would provide a more solid academic balance.

Response 12: Thank you for this important observation. As the reviewer astutely observes, our work does have limitations.

 Our revisions are as follows:

“4. Conclusions”, p.16 (line 629–635) in the revised paper.

Though the experimental validations in Section 3 have validated Dyn-Pri’s functionality and performance, there still exists some limitations to be addressed in the future research, including how to perform incremental privacy-sensitivity assessments that capture both the commonalities and the distinct requirements of diverse V2G services, and how to remain fully aligned with current vehicle-to-charger communication standards and their evolving data-collection capabilities. Of course, the mentioned issues will be the primary focus of our forthcoming Dyn-Pri+ research.

 

Round 2

Reviewer 5 Report

Comments and Suggestions for Authors

All observations have been appropriately addressed by the authors and there are no other issues to clarify.