A Novel Energy Control Digital Twin System with a Resource-Aware Optimal Forecasting Model Selection Scheme

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The paper lacks a clear comparative analysis demonstrating how the proposed control method differs from existing approaches, which diminishes the novelty claim.

2. The criteria for similarity detection in the model selection process are not sufficiently detailed, making it difficult to understand how statistical features influence model performance.

3. There is no comprehensive evaluation against state-of-the-art methods, which would provide essential context for the performance claims made in the paper.

4. The handling of dynamic resource constraints during operation is not well-explained, and the paper does not provide case studies to illustrate this adaptability.

5. The discussion on the trade-off between model complexity and forecasting accuracy is insufficient, lacking details on how this trade-off is quantified and managed in practice.

6. The experiments are limited to LPG consumption data from a specific region, and the paper does not address the generalizability of the findings to other energy domains.

7. The training process for the meta-learner and its ability to accurately reflect historical performance in model selection is not adequately discussed, leaving potential limitations unaddressed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors, regarding the submitted manuscript, the main points I critique are related to the proposal (which is unclear) and the organization of the text. Below, I list these issues point by point:

1. Section I: The manuscript's proposals/contributions, when compared to the literature review, end up getting lost within the text. It is advisable to include a table that shows the characteristics of the works in the current literature and contrasts them with the manuscript’s proposals;

2. Section II seems out of place in the manuscript. Consider merging it into Section I, as it still deals with the literature review;

3. Better explain the contribution of the work. Considering the text in lines 17–19, 73–77, 93–95, and 119–121, in addition to being somewhat repetitive, the proposal is a bit confusing and the actual contribution/proposal is unclear;

4. Lines 137–138: "The system preprocesses the measurement data and store the data into the data store" is redundant — rewrite it for clarity;

5. Line 141: clarify how the system performs the selection optimal model based on current available computing resources of the control system—  it's unclear how the selection mechanism based on available computing resources would work in this case;

6. Lines 146–147: mention which attributes were used for the combination;

7. Redundant phrase: "The energy sensors collect the data from the sensor-installed sites";

8. Figure 2 shows measurements from three locations, while the text on line 179 says it is from two locations. The chart title also seems odd — if it is an LPG sensor, why is it called an anomaly sensor?

9. (1) is an equation, not an algorithm;

10. Line 208: shouldn't it be Table 1? And line 235: Table 2?

11. There are missing figure references and some are out of order — line 196 mentions Figure 3(a), then it jumps from Figure 4 to Figure 6;

12. How was the execution time of the models measured? During these measurements, could other processes related to the operating system have affected the results?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The observations made in the first review were met.

There are still minor adjustments such as:
- hyperlinks in references
- hyperlinks in tables, figures
These hyperlinks make it easier to read and locate the text

Author Response

Comments 1: The observations made in the first review were met.
There are still minor adjustments such as:
- hyperlinks in references
- hyperlinks in tables, figures
These hyperlinks make it easier to read and locate the text

Response 1: Thank you for pointing this out. I agree with this comment. Therefore, we have added hyperlinks in references, figures and tables. The changes are colored in blue in the manuscript.
Also, I have a change on Abstract which is the latency reduction performance, and we have changed the performance to 19 times on the last sentence of the abstract.