Model Sharing and Scalability in the Real-Time Simulation and Intelligent Hierarchical Control of Discrete-Event Systems

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

The paper introduces a modelling paradigm to resolve the scalability issue stemming from the model explosion problem. It presents a method where simulation models are automatically generated by repurposing control models, intending to streamline the process of incorporating intelligence. The author asserts that this approach will diminish the modelling effort yet fails to elaborate on how it achieves a simulation modelling effort close to zero (P = 0).

According to the author, the proposed modelling paradigm addresses the challenges of providing real-time simulation models for decision-making without significant modelling effort, making the approach scalable to arbitrarily large systems. However, it lacks exposing some technical details such as:

·       How does the model explosion problem impact the scalability of distributed intelligent controllers in large automated systems?

·       What examples of large-scale automated systems could benefit from the concepts discussed in this paper?

·       How can current modelling technology be improved to address the challenges of providing intelligence to controllers in the control hierarchy?

The author establishes a theoretical basis for the proposed modelling paradigm by delineating challenges associated with traditional approaches, which involve estimating simulation models based on control model operations. However, the paper requires revision, particularly in Sections 2 and 3.

Section 2 is disproportionately lengthy and lacks emphasis on pertinent background and theoretical aspects concerning model sharing, efficiency, scalability, and how the proposed approach addresses challenges related to real-time simulation models with intelligent hierarchical control in large automated systems. It's uncertain if all the information effectively contributes to understanding the advantages of the proposed model. It is hard to establish a correlation between what is presented in Line 387-564 and the example given in L565, namely the relevance of Fig. 13 and 14?

·       Resume the assumption to the essential;

·       Fig. 1, 2 and 3 could be merged into a single presentation and resume what is relevant to retain;

·       Fig. 4, please review to provide a more contextualised example, for instance, to match with the remarks in the observation 1 and 2

·       Our Solution (Line 387) requires a deep review to focus on what is relevant. Fig. 6 is too abstract, with little or no helpful information.

·       Fig.7 – 9 + Fig. 12 -  consider reviewing how to present the algorithm decision process

Section 3: Results, draws from an illustrative example that warrants review; it may be beneficial to integrate this section with Section 4: Discussion.

 

Section 5: Conclusion lacks a concise and practical summary regarding the proposed modelling paradigm. It is excessively lengthy and lacks focus on pertinent aspects, failing to summarise the theoretical foundation supporting the proposed approach briefly. Key findings should be provided comparing the manual creation of simulation models versus a (semi)automated approach to modelling/integrating intelligence across all control hierarchy levels.

Comments on the Quality of English Language

Minor English language editing is required, but most importantly, sections 2 and 3 should be resumed to what is relevant.

Author Response

The paper introduces a modelling paradigm to resolve the scalability issue stemming from the model explosion problem. It presents a method where simulation models are automatically generated by repurposing control models, intending to streamline the process of incorporating intelligence. The author asserts that this approach will diminish the modelling effort yet fails to elaborate on how it achieves a simulation modelling effort close to zero (P = 0).

Response: The paper presents a tool that allows the simulation models to be created by recycling existing control models. This results in P = 0. In lines 758 to 781 I added a section discussing scalability and what it means. Its true that some effort is needed in creating the simulation models but its does not grow exponentially. This is discussed in this added material.

According to the author, the proposed modelling paradigm addresses the challenges of providing real-time simulation models for decision-making without significant modelling effort, making the approach scalable to arbitrarily large systems. However, it lacks exposing some technical details such as:

• How does the model explosion problem impact the scalability of distributed intelligent controllers in large automated systems?

Response: Section 3 Materials and Methods (old section 2) is entirely dedicated to the impact the model explosion problem has on scalability. That is the whole purpose of the mathematical analysis shown.

• What examples of large-scale automated systems could benefit from the concepts discussed in this paper?

Response: That is shown in the introduction. The smart grid, large transportation systems etc. See the first paragraph of the introduction.

• How can current modelling technology be improved to address the challenges of providing intelligence to controllers in the control hierarchy?

Response: I do not know. This paper is not about how the current technology can be improved but rather how we can use a new technology to accomplish this.

The author establishes a theoretical basis for the proposed modelling paradigm by delineating challenges associated with traditional approaches, which involve estimating simulation models based on control model operations. However, the paper requires revision, particularly in Sections 2 and 3.

Response: That is not correct. The work I show in estimating simulation models based on control model operations is in the literature review. It does not represent the traditional approaches but rather new approaches competing with the presented one. Our presented approach is not estimating the model from the control models but rather recycling them.

Unfortunately, your request is too vague for me to improve. This paper is in its 7^th revision and currently 3 of the 4 reviewers are suggesting only very specific changes which I addressed. Please provide specific changes you will like to see.

Section 2 is disproportionately lengthy and lacks emphasis on pertinent background and theoretical aspects concerning model sharing, efficiency, scalability, and how the proposed approach addresses challenges related to real-time simulation models with intelligent hierarchical control in large automated systems. It's uncertain if all the information effectively contributes to understanding the advantages of the proposed model. It is hard to establish a correlation between what is presented in Line 387-564 and the example given in L565, namely the relevance of Fig. 13 and 14?

Response: Section 2 (now 3) is long since its complex and involves mathematical proofs and derivations. The whole paper is 24 pages which is over by 4 pages. I believe the paper will become 20 pages when it’s put into the final version formatting. I removed lots of text in the last revision to bring the paper down to its current size. I do not believe there is still material that is not relevant. At this time removing material may lead to problems with the other reviewers that already accepted this section. The example uses the tool developed with the presented solution to show it works and does in fact require no modeling effort. I do not understand how that is not correlated. The figures 13 and 14 are not required but give the reader a sense that the system being controlled is a real physical system and also helps provide context to the size and complexity of the system in the example. Therefore, I left it in the paper.

• Resume the assumption to the essential;

Response: I left the assumptions as a table to emphasize them. They used to be integrated into the text and reviewers indicated they were confused. Placing them into a table made them stand out and help with confusion. All assumptions are needed to be understood as I found they are not obvious to reviewers.

• Fig. 1, 2 and 3 could be merged into a single presentation and resume what is relevant to retain;

Response: They were merge into a single figure. I do not believe this helps but it may reduce the size of the paper. All parts of those figures are relevant and necessary.

• Fig. 4, please review to provide a more contextualised example, for instance, to match with the remarks in the observation 1 and 2

Response: I do not understand this comment. Figure 4 expresses exactly what I need for my derivations which lead to observations 1 and 2.

• Our Solution (Line 387) requires a deep review to focus on what is relevant. Fig. 6 is too abstract, with little or no helpful information.

Response: I did not find any material in the solution that is not relevant. Fig 6 is actually the most important figure in the paper and summarises the complete solution. It has lots of information in it. Once the solution is better understood, the figure will be very helpful in providing a unified and summarized view of the solution.

• Fig.7 – 9 + Fig. 12 -  consider reviewing how to present the algorithm decision process

Response: I am not sure of the format you are asking for. Best practice to present algorithms is to use pseudo code as is done here. I did enhance the loops in the pseudo code to emphasize them as well as adding bold face to key words and more clearly indicating the end of if statements and loops.

Section 3: Results, draws from an illustrative example that warrants review; it may be beneficial to integrate this section with Section 4: Discussion.

Response: I agree but this section was once integrated with the Results section, but a reviewer requested I separate them. I did make changes to both sections based on the recommendations of the other reviewers.

 Section 5: Conclusion lacks a concise and practical summary regarding the proposed modelling paradigm. It is excessively lengthy and lacks focus on pertinent aspects, failing to summarise the theoretical foundation supporting the proposed approach briefly. Key findings should be provided comparing the manual creation of simulation models versus a (semi)automated approach to modelling/integrating intelligence across all control hierarchy levels.

Response: I reread the conclusions many times and I do not know how I can better summarize the work. The first paragraph summarizes the theory. The next paragraph summarizes the model explosion problem and shows the key finding, p is proportional to L. The next paragraph shows some key details related to the implementation of the solution. Then finally it summarizes the example followed by future work. It’s basically a summarized version of the whole paper referencing key findings. Not sure how else to improve.

Minor English language editing is required, but most importantly, sections 2 and 3 should be resumed to what is relevant.

Response: I fixed many grammatical errors.

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The manuscript has undergone a thorough revision, and several concerns raised by the reviewers and the editor have been addressed. Overall, the paper has improved, but in my opinion, there are still some aspects that need to be revised."

 

The author shortened the literature review, also removing its label. Generally, I disagree with such an approach. A journal paper should have archival value, and in this respect, an in-depth discussion of the state-of-the-art, which clearly outlines the context and compares the proposed approach with existing ones, is somehow needed. In this respect, the editor's comments do not suggest removing the related work section but simply moving it away from the introduction. Therefore, I suggest adding a "Related work" section which should also be extended with a more technical comparison of your work with existing similar ones, especially those in the field of model composition.

 

The discussion on the scalability of the proposed approach (and consequently its use in a real-time domain) is still not entirely convincing. The paper states (lines 402-404) that “our solution is to include all the individual control models along with a model of the communication system into a new single model that can then be used in a simulation following well-known algorithms for discrete-event simulation.” Moreover, at line 374 it is also stated that “One million controllers are not unrealistic in future systems such as the Smart Grid we envision”. In this respect, while the proposed approach solves the state explosion problem by reusing control models, this work neither discusses nor provides any concrete elements to illustrate i) how to manage the resources needed to create a simulation model consisting of a very large number of models (as in the cited case), and ii) execute it on a single computing node.

 

This is a crucial aspect as the proposed approach aims at address real time systems.

 

In this respect, the appreciable revision of the example application does not help to understand these aspect as it does not provide any performance-related element. Therefore, I suggest to further revise the example, so to add a discussion about time- and performance- related aspects.

 

MINOR ISSUES:

-       In general, I would suggest revising the English;

-       line 98: review the punctuation;

-       Figure 13 (caption): remove the bold typeface;

-       Figure 7 and 15: check the page layout (position of the caption);

Comments on the Quality of English Language

See comments for Authors.

Author Response

The manuscript has undergone a thorough revision, and several concerns raised by the reviewers and the editor have been addressed. Overall, the paper has improved, but in my opinion, there are still some aspects that need to be revised."

 

The author shortened the literature review, also removing its label. Generally, I disagree with such an approach. A journal paper should have archival value, and in this respect, an in-depth discussion of the state-of-the-art, which clearly outlines the context and compares the proposed approach with existing ones, is somehow needed. In this respect, the editor's comments do not suggest removing the related work section but simply moving it away from the introduction. Therefore, I suggest adding a "Related work" section which should also be extended with a more technical comparison of your work with existing similar ones, especially those in the field of model composition.

 

Response: The references removed were at the request of a reviewer. I tried to remove references that gave background information only and did not present work directly related so I believe it did not hurt. At that time I added new references more closely related.

 

I moved the Related Work section from the Introduction and placed it as its own section.

See lines 145 to 150 and 175 to 180 for new references. There is no other work related to the recycling of models or related to reducing the simulation modeling effort in an intelligent control setting. We already presented this work in lines 151 to 161. No other work directly related was found.

 

The discussion on the scalability of the proposed approach (and consequently its use in a real-time domain) is still not entirely convincing. The paper states (lines 402-404) that “our solution is to include all the individual control models along with a model of the communication system into a new single model that can then be used in a simulation following well-known algorithms for discrete-event simulation.” Moreover, at line 374 it is also stated that “One million controllers are not unrealistic in future systems such as the Smart Grid we envision”. In this respect, while the proposed approach solves the state explosion problem by reusing control models, this work neither discusses nor provides any concrete elements to illustrate i) how to manage the resources needed to create a simulation model consisting of a very large number of models (as in the cited case), and ii) execute it on a single computing node.

 

Response: In lines 758 to 781 I added a discussion of scalability and what it means. I addressed the creation of models for very large-scale systems as well as running those models on a single computer.

 

This is a crucial aspect as the proposed approach aims at address real time systems. In this respect, the appreciable revision of the example application does not help to understand these aspect as it does not provide any performance-related element. Therefore, I suggest to further revise the example, so to add a discussion about time- and performance- related aspects.

 

Response: Lines 758 to 781 were added to discuss the running times and performance aspects. Recall this paper and the presented methodology is about building simulation models with less effort not necessarily building better or more efficient models. I just need to show that the resultant simulation models are equivalent or just as good to the manually created ones. How they are run, how fast, what they do etc are outside the scope of this work provided that they are the same as the models created using traditional modeling tools. In the lines indicated above I did however discuss some of these issues that apply to all simulation models regardless of how they were created.

 

MINOR ISSUES:

-       In general, I would suggest revising the English;

Response: Done.

-       line 98: review the punctuation;

Response: Added commas.

-       Figure 13 (caption): remove the bold typeface;

Response: Done

-       Figure 7 and 15: check the page layout (position of the caption);

Response: The formatting of the figures and tables is dependent on the page breaks and until the final version that is not in the review style, the formatting cannot be accomplished. Formatting it now will only require change once the paper is put into the publication format. But I will definitely reformat the whole paper for the final version. This version has a large margin to the left. Once in the final margins everything will move.

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

I would recommend the authors to add  more references and more recent references. 1/4 of the references represent self-citations.

How do the authors plan to evaluate their theoretical algorithm with a simulation, and how they can test it with a large scale system?

Comments on the Quality of English Language

The authors need to correct some grammatic errors in the text, improve the formatting of the figures and the formulas, also the text.

Author Response

I would recommend the authors to add  more references and more recent references. 1/4 of the references represent self-citations.

Response: See lines 145 to 150 and 175 to 180. There is no other work related to the recycling of models or related to reducing the simulation modeling effort in an intelligent control setting. We already presented this work in lines 151 to 161. No other work directly related was found.

How do the authors plan to evaluate their theoretical algorithm with a simulation, and how they can test it with a large scale system?

Response: The methodology cannot be tested on a real large system since it is too expensive to borrow a real system for this testing. The example is on a small but real physical system with real hardware so it’s the closest thing we can use. Text was added to address this concern see lines 684 to 695.

The authors need to correct some grammatic errors in the text, improve the formatting of the figures and the formulas, also the text.

Response: I gave the paper a total review looking for grammatical error and I found and fix many. The formatting of the figures and tables is dependent on the page breaks and until the final version that is not in the review style, the formatting cannot be accomplished. Formatting it now will only require change once the paper is put into the publication format. But I will definitely reformat the whole paper for the final version. This version has a large margin to the left. Once in the  final margins everything will move.

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

The paper describes an study about simulation techniques for event-orientes systems. The proposal is technically sound in general terms.

The claims in the intriduction are justified and the problem is correctly described. Description are OK. I would not say they are exhaustive, but figures are provided, and that's could be enough considering the topic and the contribution. I think this is enough to ensure future replicability and readability. 

References are timely and the topic mtaches the scope of the journal. 

Results are based on individual uses cases, and are also coherent with the state of the art and validate the initial hypothes. However, in my opinion, some improvements need to be done to this section:

1) First, you need to describe your experimental setup with details. Software and hardware you used. How result were calculated. Why these uses cases? Which are the main scenarios you are covering with this new approach?

2) As all you results are simulations and individual use cases you need to discuss the validity threats, at least the internal an external. 

Author Response

The paper describes an study about simulation techniques for event-orientes systems. The proposal is technically sound in general terms.

Response: Yes. Thank you

The claims in the intriduction are justified and the problem is correctly described. Description are OK. I would not say they are exhaustive, but figures are provided, and that's could be enough considering the topic and the contribution. I think this is enough to ensure future replicability and readability.

Response: I agree. 

References are timely and the topic mtaches the scope of the journal. 

Response: I agree.

Results are based on individual uses cases, and are also coherent with the state of the art and validate the initial hypothes. However, in my opinion, some improvements need to be done to this section:

1) First, you need to describe your experimental setup with details. Software and hardware you used. How result were calculated. Why these uses cases? Which are the main scenarios you are covering with this new approach?

Response: In lines 684 to 695 I added some descriptions of the set up I used and the scenarios I used as well. I showed why these scenarios were selected and how the results were verified.

2) As all you results are simulations and individual use cases you need to discuss the validity threats, at least the internal an external. 

Response: The paper is about building simulation models with less effort and no so much about the benefits of using them. In lines 696 to 706 I discussed two scenarios that are common.  

Round 2

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

The paper introduces a modeling paradigm to resolve the scalability issue stemming from the model explosion problem. It presents a method where simulation models are automatically generated by repurposing control models, intending to streamline the process of incorporating intelligence. The author asserts that this approach will diminish the modeling effort yet fails to elaborate on how it achieves a simulation modeling effort close to zero (P = 0).

According to the author, the proposed modeling paradigm addresses the challenges of providing real-time simulation models for decision-making without significant modeling effort, making the approach scalable to arbitrarily large systems. A minor improvement should be considered to clarify the research's main findings in the Discussion section or in the Conclusion section, namely:

·       How does the model explosion problem impact the scalability of distributed intelligent controllers in large automated systems?

·       What examples of large-scale automated systems could benefit from the concepts discussed in this paper?

 

·       How can current modeling technology be improved to address the challenges of providing intelligence to controllers in the control hierarchy?

Author Response

Comment 1:

The paper introduces a modeling paradigm to resolve the scalability issue stemming from the model explosion problem. It presents a method where simulation models are automatically generated by repurposing control models, intending to streamline the process of incorporating intelligence. The author asserts that this approach will diminish the modeling effort yet fails to elaborate on how it achieves a simulation modeling effort close to zero (P = 0).

According to the author, the proposed modeling paradigm addresses the challenges of providing real-time simulation models for decision-making without significant modeling effort, making the approach scalable to arbitrarily large systems. A minor improvement should be considered to clarify the research's main findings in the Discussion section or in the Conclusion section, namely:

• How does the model explosion problem impact the scalability of distributed intelligent controllers in large automated systems?
• What examples of large-scale automated systems could benefit from the concepts discussed in this paper?

Response 1:

I added lines 739 to 756. I added clarification on the research’s two main goals, addressed the impact the model explosion problem has on the scalability of the controller, and gave examples of the large-scale distributed systems that can benefit from the proposed research.

Comment 2:

How can current modeling technology be improved to address the challenges of providing intelligence to controllers in the control hierarchy?

Response 2:

I do not understand this comment. I created this new modelling technology because I could not figure out how to address this problem using the current technology. The current modelling technology does not support model sharing. So, to answer your question, I do not think the current modelling technology can be improved to address this modelling challenge. That is why I created this new technology. I hope I answered your question.

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The author has properly addressed the suggestions I proposed in my review. In this latest revision, the paper has been considerably enhanced and, in my opinion, it can be considered for publication.

Author Response

Comment 1:

The author has properly addressed the suggestions I proposed in my review. In this latest revision, the paper has been considerably enhanced and, in my opinion, it can be considered for publication.

Response 1:

Thank you.

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

Comments and Suggestions for Authors

This paper introduces an approach for implementing intelligent controller in arbitrarily large automated systems. It focuses on the scalability of hierarchical software controllers, specifically regarding the availability of intelligent decision-making capability.

 

In the proposed approach, the various control nodes implement relevant simulation systems that provide the required decision-making capabilities.

 

The paper addresses a relevant topic that fits the scope of the journal well. The inherent complexity of a large-scale automated system, its size, the number of components and relevant interconnection channels and, finally, the time-step duration required for synchronizing the simulation logical time with the real time, might lead to a model explosion problem and dictates the computing power required.

 

Unfortunately, in my opinion, this paper shows some weaknesses.

 

1) INTRODUCTION

Lines 63-68: It is illustrated what is claimed as the (most used) existing approach for building intelligence in a controller: “to implement intelligence in a controller, a simulation model must be created for that controller that encompasses the logic in that particular controller’s control domain”. To enforce the motivation of the paper, I suggest including in this section some references from the literature to existing approaches and anticipate why the method proposed in this paper (model composition) is not currently adopted.

 

2)RELATED WORK

In my opinion, the proposed literature review misses some relevant topics which are at the basis of this work: methods for model composition (specifically in the case of simulation models), approaches for distributed and parallel simulation, and specifically limitations and existing solution for the (real-time) simulation of automated and large systems (e.g., smart grid). As valuable contributions are available in literate, I would appreciate a discussion which also clearly compares existing approaches with the proposed one.

 

3) GENERAL TREE

In my opinion, too much room is spent for describing the model explosion problem. I suggest reducing this section to save space for better addressing other proposed comments.

 

4) PROPOSED SOLUTION

 

Pag 18: In a simulation system, the list of upcoming events is often denoted as a “pending” event list. The use of scheduled/pending event list in lines 696-713 might be misleading, therefore I suggest better distinguishing the terms used for denoting simulated events and real events. Similarly, the two labels in Figure 6 (Event List and pending Event List) should be revised accordingly.

 

Lines 676 – 677: “Since we are using the same model source code used for control, all its logic, including any errors, will appear in the simulate model.”. Does not the use of a control model as a part of the simulation systems require any adaptation? Please clarify this aspect.

 

Lines 677-821: My main concern is related to the soundness of the proposed solution and its actual scalability. Maybe I’m missing something, but I can’t find any novel contribution/approach in what is stated in this paragraph that includes many well-known ideas, methods and approaches commonly addressed or adopted when it comes to deal with parallel, distributed and local simulation.

 

As an example, lines 718-722 simply discus model composition issues, while Figures 7, 8 and 9 depict algorithms commonly adopted for managing events in a simulation.

 

 

Basically, the proposed modeling approach consists in creating a simulation model by composing the various existing models in a single program. Moreover, a global event list is introduced that stores a copy of every event in every controller object.

 

The novelty and the effectiveness of such an approach need to be better discussed, also considering existing literature as previously suggested.

 

Specifically:

 

A) SCALABILITY: As stated in Section 1, this work addresses arbitrarily large systems, which might generally consist of a huge number of components, and which require a simulation program composed of as many control models as the real system has. Doesn’t such an approach have any scalability issue to be investigated? An appropriate discussion should be provided.

 

 B) NOVELTY: To better convince readers about the novelty of the proposed approach, it should be compared with existing literature in the field of parallel/distributed simulation.

 

 

5) EXMPLE APPLICATION

To demonstrate the validity, feasibility, and effectiveness of the proposed approach, I suggest to completely revise this Section. In my opinion, the experimentation discussion should focus on a single case that will show in detail (and clarify) how concrete aspects are actually handled, rather than discussing more cases but from a general point of view. Specifically, this section should better clarify the following aspects:

- how the reuse of controller modes in the simulator is actually carried out;

- the value of P obtained in the example;

- the performance of the control system.

 

Comments on the Quality of English Language

This paper introduces an approach for implementing intelligent controller in arbitrarily large automated systems. It focuses on the scalability of hierarchical software controllers, specifically regarding the availability of intelligent decision-making capability.

 

In the proposed approach, the various control nodes implement relevant simulation systems that provide the required decision-making capabilities.

 

The paper addresses a relevant topic that fits the scope of the journal well. The inherent complexity of a large-scale automated system, its size, the number of components and relevant interconnection channels and, finally, the time-step duration required for synchronizing the simulation logical time with the real time, might lead to a model explosion problem and dictates the computing power required.

 

Unfortunately, in my opinion, this paper shows some weaknesses.

 

1) INTRODUCTION

Lines 63-68: It is illustrated what is claimed as the (most used) existing approach for building intelligence in a controller: “to implement intelligence in a controller, a simulation model must be created for that controller that encompasses the logic in that particular controller’s control domain”. To enforce the motivation of the paper, I suggest including in this section some references from the literature to existing approaches and anticipate why the method proposed in this paper (model composition) is not currently adopted.

 

2)RELATED WORK

In my opinion, the proposed literature review misses some relevant topics which are at the basis of this work: methods for model composition (specifically in the case of simulation models), approaches for distributed and parallel simulation, and specifically limitations and existing solution for the (real-time) simulation of automated and large systems (e.g., smart grid). As valuable contributions are available in literate, I would appreciate a discussion which also clearly compares existing approaches with the proposed one.

 

3) GENERAL TREE

In my opinion, too much room is spent for describing the model explosion problem. I suggest reducing this section to save space for better addressing other proposed comments.

 

4) PROPOSED SOLUTION

 

Pag 18: In a simulation system, the list of upcoming events is often denoted as a “pending” event list. The use of scheduled/pending event list in lines 696-713 might be misleading, therefore I suggest better distinguishing the terms used for denoting simulated events and real events. Similarly, the two labels in Figure 6 (Event List and pending Event List) should be revised accordingly.

 

Lines 676 – 677: “Since we are using the same model source code used for control, all its logic, including any errors, will appear in the simulate model.”. Does not the use of a control model as a part of the simulation systems require any adaptation? Please clarify this aspect.

 

Lines 677-821: My main concern is related to the soundness of the proposed solution and its actual scalability. Maybe I’m missing something, but I can’t find any novel contribution/approach in what is stated in this paragraph that includes many well-known ideas, methods and approaches commonly addressed or adopted when it comes to deal with parallel, distributed and local simulation.

 

As an example, lines 718-722 simply discus model composition issues, while Figures 7, 8 and 9 depict algorithms commonly adopted for managing events in a simulation.

 

 

Basically, the proposed modeling approach consists in creating a simulation model by composing the various existing models in a single program. Moreover, a global event list is introduced that stores a copy of every event in every controller object.

 

The novelty and the effectiveness of such an approach need to be better discussed, also considering existing literature as previously suggested.

 

Specifically:

 

A) SCALABILITY: As stated in Section 1, this work addresses arbitrarily large systems, which might generally consist of a huge number of components, and which require a simulation program composed of as many control models as the real system has. Doesn’t such an approach have any scalability issue to be investigated? An appropriate discussion should be provided.

 

 B) NOVELTY: To better convince readers about the novelty of the proposed approach, it should be compared with existing literature in the field of parallel/distributed simulation.

 

 

5) EXMPLE APPLICATION

To demonstrate the validity, feasibility, and effectiveness of the proposed approach, I suggest to completely revise this Section. In my opinion, the experimentation discussion should focus on a single case that will show in detail (and clarify) how concrete aspects are actually handled, rather than discussing more cases but from a general point of view. Specifically, this section should better clarify the following aspects:

- how the reuse of controller modes in the simulator is actually carried out;

- the value of P obtained in the example;

- the performance of the control system.

 

Reviewer 2 Report

Comments and Suggestions for Authors

This document spans 34 pages, featuring an extensive section on "2. Materials and Methods" (comprising 18 pages). However, it's noted that the presentation of the results lacks a methodological framework that promotes the replicability of the experiment's findings. Furthermore, the author's ability to concisely contextualize and engage the reader within a max of 20 pages is vital in assessing the quality and maturity of a scientific roadmap. This should involve a clear exposition of the research's added value, an objective and formal presentation of significant achievements, and an explanation of the methodology(ies) employed (covered in the Discussion section). In the Conclusion section, it's expected that the main findings are succinctly summarized, and insights for future research are provided. The author is encouraged to revise the paper, condensing the content to fit within a maximum of 20 pages and outline the core features that are determinants from an evidence-base perspective.

Line 1012:  Discussion and Conclusion are two separate sections each one with specific and distinct purposes, not possible to be merged

Line: 93 + 99:  “This work may be a bit ahead of its time…” + “…the problem this paper is addressing does not necessarily exist today but are in the near future…” by definition, a research work outlines some innovation that tries to provide a contribution to solving problems or challenges for which there is not yet a well-known solution.  

 Line 765 – 780: Fig7 until Fig. 9, a strong recommendation to adopt BPMN as a standard set of diagramming conventions that can be used to map simulation flows (provides more information than the flow chart), removing any ambiguity or interpretation subjectivity which also eliminates the need to complement the data sequence flow with a pseudo code.

 Line 1048: References, need to uniform the way DOI are presented; some references are not well referenced (e.g., [22] the URL points to the https://www.proquest.com/ website?)

 

Comments on the Quality of English Language

Minor editing of the English language required