An Optimization Method for Collaborative Radar Antijamming Based on Multi-Agent Reinforcement Learning
Round 1
Reviewer 1 Report
In this paper, a radar collaborative antijamming method is proposed based upon the multi-agent reinforcement learning to improve the detection performance. Simulation results are provided to verify the effectiveness of the proposed algorithm. Overall, this is a good work. Some suggestions are given below.
1. It's better to reorganize the introduction section to make it more logical.
2. Please provide a flowchart of the proposed algorithm.
3. More comparison results on the different jammer can be added.
4. It's better to point out the potential limitation of this work in the conclusion part.
None
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The article proposes a methodology based on reinforcement learning for multi-agent systems to address the antijamming problem from a collaborative point of view, the problem is treated as a zero-sum game. Simulation results are presented that support the obtained results. The article is interesting and contributes to knowledge, here are some comments to improve it.
1. the introduction provides adequate preliminary information on the problem to be addressed, however, it is recommended to add appointments with more current products, preferably from 2022 or 2023
2. Why do the authors assume that the total jamming capability of the shipborne jammer is constant and limited?, What would happen if it were not bounded? What modifications would have to be made to the proposed methodology?
3. in the matrix of expression 1, is m=n assumed?
4. in the approach used by the authors of a completely competitive zero-sum game, it implies that it is a static game with no state signal and therefore there is no system dynamics. How could the authors justify that this is the best way to model the problem to be solved?
5. it is recommended that the quality of figure 3a be improved, so that the data can be better appreciated.
6. In the conclusion, it is recommended that the authors expand the discussion about the results obtained against those found in the literature.
It is recommended to review minor grammatical errors in the text, in general what is written is understandable
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Dear authors,
Thank you for this interesting paper. It is an interesting problem which has to be solved. From my point of view, the paper is not balanced well: You have 3.5 pages of introduction, eight pages of methods, two pages of results and a 0.5-page discussion & conclusion. You have to extend the results and discussion so that the readers can clearly assess your contribution and see its advantages. Currently, you are not mentioning how well it works regarding percentages (e.g. about 70% winning probability with MADDGP vs less than 50% for the other cases). Also, I am missing an explanation and/or outlook regarding the limitations of your simulations (e.g. you did not consider the direction of the attack as well as the beamwidth of the jammer. Likely all attacking missiles would be coming from the same/near-by launch site/s and therefore be jammed at the same time). Please extend and explain better.
Thank you very much, and all the best.
Comments for author File: 
Comments.pdf
Please review the grammar. I mentioned some errors in the beginning, but there are more in the following paragraphs. Have a focus on singular/plural, word combinations, commas, font types (text vs equation) etc.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The authors have addressed all the questions.
None
Reviewer 3 Report
Thanks for your reply and the extension of the discussion!
Only minor issues.
