An Anti-Jamming Hierarchical Optimization Approach in Relay Communication System via Stackelberg Game

Round 1

Reviewer 1 Report

The authors formulate a communication problem involving relays and a jammer. The paper follows along the lines of game theory applied to communication problems.  One concern that the reviewer had is that the authors only seem to be aware of new, minor contributions to the field and isn't aware of work from the 1990s and early 2000s. 

For example, there are numerous papers on studying relays using game theory. For a good reference there, consider work by KJR Liu's group (this book might be helpful: https://www.cambridge.org/gb/academic/subjects/engineering/wireless-communications/cognitive-radio-networking-and-security-game-theoretic-view?format=HB&isbn=9780521762311 )

Also, there are many works from the Inria group (Eitan Altman) on jamming and communication modeling with game theory.

The reviewer would also suggest looking at papers by Garnaev and Trappe on jamming and anti-jamming. Particularly, the use of Stackelberg games by those authors might be relevant. 

Likewise work by R. Poovendran's group at the University of Washington also is relevant to consider. 

Generally, the reviewer feels that the authors need to broaden their literature survey. Additionally, the authors then need to make more comparison with prior art. What are they advancing?

What does "Considering the influence of distance because of channel fading simultaneously" mean? Are the authors trying to say "Considering both the influence of distance and channel fading,..."?

Don't use "RSA" for random selection algorithm. RSA is sort of a reserved word/abbreviation in the security community.  How about just "RS"?

In general, the paper is well written but has grammatical and word choice mistakes. That said, the organization of the paper is good. Some examples of problems with the paper's writing include:

1)  The authors frequently miss the word "the" and other articles, which are needed. For example, "When jammer senses" should be "When the jammer senses". This type of mistake is minor but occurs frequently.

2) The word "updation" is not a real word. 

The authors should go through and check the grammar and word choice in the paper. A little editing effort (perhaps from grammarly or an online service) would be a low cost way to improve the readability of the paper.  

Author Response

Response to Reviewer 1 Comments

 

First of all, we would like to express our great thanks to the reviewer-1’s helpful suggestions and valuable comments. We have carefully conducted the reviewer-1’s comments and the detailed point-by-point response to the comments is given below and the revised portions in the revised manuscript are marked in red. We hope that the revision will be found satisfactory and this paper is now suitable for publication.

 

Point 1:

The authors formulate a communication problem involving relays and a jammer. The paper follows along the lines of game theory applied to communication problems.  One concern that the reviewer had is that the authors only seem to be aware of new, minor contributions to the field and isn't aware of work from the 1990s and early 2000s. 

 

For example, there are numerous papers on studying relays using game theory. For a good reference there, consider work by KJR Liu's group (this book might be helpful: https://www.cambridge.org/gb/academic/subjects/engineering/wireless-communications/cognitive-radio-networking-and-security-game-theoretic-view?format=HB&isbn=9780521762311 )

 

Also, there are many works from the Inria group (Eitan Altman) on jamming and communication modeling with game theory.

 

The reviewer would also suggest looking at papers by Garnaev and Trappe on jamming and anti-jamming. Particularly, the use of Stackelberg games by those authors might be relevant. 

 

Likewise work by R. Poovendran's group at the University of Washington also is relevant to consider. 

 

Generally, the reviewer feels that the authors need to broaden their literature survey. Additionally, the authors then need to make more comparison with prior art. What are they advancing?

 

Response 1:

We are very grateful for your helpful comments. We have added some references in the revised version to broaden our literature survey, and the relative discusses are given as follows:

In [R.1], the authors focused on game-theoretical approaches in cognitive communication and networking systems, in which a stochastic anti-jamming game was proposed to design the optimal adaptive defense strategies against cognitive malicious attackers. In [R.2], the authors designed an opportunistic wireless communication setting and modelled the asset selling problem as a game theoretic variant in the completely observable and the partially observable cases respectively. In [R.3], the authors studied the problem of resource assigning in a single carrier communication system in the presence of a jammer, a Bayesian jamming game framework was proposed and the Nash strategy was also compared to the Stackelberg strategy to verify its sensibility. In [R.4], the authors studied a new anti-jamming problem of unknown nodes in a peer-to-peer network attacked by a random jammer or an intelligent one. In [R.5], considering the idealized case and potential energy constraint, the authors investigated the jamming attacks optimization problem that jammer could control the probability of jamming and transmission range to cause maximal damage to wireless network. Taking the perspective of Stackelberg game, the authors combed the anti-jamming technologies under different scenarios in [R.6], moreover, an anti-jamming decision-making framework was also proposed based on the adversarial characteristics between user and jammer, incomplete information and so on.

The existed works investigated the anti-jamming problems from the perspectives of adaptive defense strategies, potential energy constraint, jamming modes, ideal and non-ideal cases in different communication scenarios to perfect the study. While game theoretical approaches, such as stochastic game and Stackelberg game, are just methods which can help us analyse the anti-jamming problems effectively. And the main difference is that, in the presence of an intelligent jammer, we investigate the relay selection problem under power energy constraint with imperfect information.

 

[R.1] Ray Liu, K. J.; Wang, B. Cognitive Radio Networking and Security: A Game-Theoretic View. Cambridge University Press. 2010.

[R.2] Naveen, K. P.; Altman, E.; Kumar, A. Competitive Selection of Ephemeral Relays in Wireless Networks. IEEE J. Sel. Areas Commun. 2017, 35, 586-600.

[R.3] Garnaev, A.; Trappe, W.; Petropulu, A. Combating Jamming in Wireless Networks: A Bayesian Game with Jammer’s Channel Uncertainty. ICASSP 2019.

[R.4] Garnaev, A.; Liu, Y.; Trappe, W. Anti-jamming Strategy Versus a Low-Power Jamming Attack When Intelligence of Adversary’s Attack Type is Unknown. IEEE Transactions on Signal and Information Processing over Networks. 2016, 2, 49-56.

[R.5] Li, M.; Koutsopoulos, I.; Poovendran, R. Optimal Jamming Attacks and Network Defense Policies in Wireless Sensor Networks. IEEE INFOCOM 2007.

[R.6] Jia, L.; Xu, Y.; Sun, Y.; Feng, S.; Anpalagan, A. Stackelberg game approaches for anti-jamming defence in wireless networks. IEEE Wireless Commun. Mag. 2018, 25, 120–128.

 

Point 2:

What does "Considering the influence of distance because of channel fading simultaneously" mean? Are the authors trying to say "Considering both the influence of distance and channel fading, ..."?

 

Response 2:

Thank you very much for your helpful comments. We have changed the previous description to “Considering the influence of distance on channel fading,...”. In the large-scale fading, the channel fading is influenced only by the distance and path-loss factor according to [R.7], in which is the channel gain. What we want to say is that the influence of distance can be viewed as the influence of channel gain .

 

[R.7] Goldsmith, A. Wireless Communications. Cambridge U.K.: Cambridge Univ. Press 2005.

 

Point 3:

Don't use "RSA" for random selection algorithm. RSA is sort of a reserved word/abbreviation in the security community.  How about just "RS"?

 

Response 3:

Thank you for your significant comments. In order to avoid the misunderstanding of random selection algorithm, we decide not to use the abbreviation “RSA” and use random selection algorithm directly.

 

Point 4:

In general, the paper is well written but has grammatical and word choice mistakes. That said, the organization of the paper is good. Some examples of problems with the paper's writing include:

 

1)  The authors frequently miss the word "the" and other articles, which are needed. For example, "When jammer senses" should be "When the jammer senses". This type of mistake is minor but occurs frequently.

 

2) The word "updation" is not a real word. 

 

Response 4:

Thanks a lot for the meaningful comments. Some needed articles are added and marked in red in the revised version. The word “updation” has been deleted. We replace the previous sentence “We consider user and jammer have different time scales of strategy updation. On the one hand, the period of jammer’s strategy updation is an epoch. On the other hand, in order to improve anti-jamming ability, user has a quicker time scale of strategy updation than jammer to adjust timely and obtain the optimal strategy” as “We consider user and jammer update their respective strategies with different time scales. On the one hand, jammer updates its power strategy every epoch. On the other hand, in order to improve anti-jamming ability, user updates its joint relay and power strategy quicker than jammer to adjust timely and obtain the optimal strategy”.

 

Point 5:

The authors should go through and check the grammar and word choice in the paper. A little editing effort (perhaps from grammarly or an online service) would be a low cost way to improve the readability of the paper.  

 

Response 5:

Thank you for your helpful comment. We have checked corrected the grammar, word choice and unclear descriptions in details. The paper have been checked out and polished to improve the quality. Hope the revised paper would be better.

 

Finally, we sincerely thank the reviewer-1’s meaningful comments and suggestions. We have tried our best to improve the manuscript and appreciate the reviewer-1’s warm work earnestly, and sincerely hope that the corrections will meet with approval.

Reviewer 2 Report

The paper presents a hierarchical joint optimization algorithm to select a relay and a transmission power in a relay communication system to improve anti-jamming ability. The problem is solved using a Q-learning algorithm and a multi-armed bandit method. The paper formulates the problem in a good way and proposes a good solution to it, however, I have only some remarks:

The related work discusses only the previous works done by the authors and not the similar works done by other authors like for example

Jia, Luliang, et al. "Stackelberg game approaches for anti-jamming defence in wireless networks." IEEE Wireless Communications 25.6 (2018): 120-128

In the title, the words “Joint Optimization Approach” give a feeling that the user and the jammer selections are optimized jointly, however they are optimized hierarchically. The joint optimization is used for selecting a relay and a transmission power from the user side. Therefore, I prefer to change the title to “Hierarchical Optimization Approach”.

As the symbols are defined in Table 1, there is no need to redefine them inside the text.

Author Response

Response to Reviewer 2 Comments

 

First of all, we would like to express our great thanks to the reviewer-2’s helpful suggestions and valuable comments. We have carefully conducted the reviewer-2’s comments and the detailed point-by-point response to the comments is given below and the revised portions in the revised manuscript are marked in red. We hope that the revision will be found satisfactory and this paper is now suitable for publication.

 

Point 1:

The paper presents a hierarchical joint optimization algorithm to select a relay and a transmission power in a relay communication system to improve anti-jamming ability. The problem is solved using a Q-learning algorithm and a multi-armed bandit method. The paper formulates the problem in a good way and proposes a good solution to it.

 

Response 1:

Thanks for your helpful comment. In the revised manuscript, we have fixed some unclear parts to improve the accuracy. Some mistakes are corrected and unclear descriptions are explained in details. The paper have been checked out and polished to improve the quality. Hope the revised paper would be better.

 

Point 2:

The related work discusses only the previous works done by the authors and not the similar works done by other authors like for example:

Jia, Luliang, et al. "Stackelberg game approaches for anti-jamming defence in wireless networks." IEEE Wireless Communications 25.6 (2018): 120-128.

 

Response 2:

Thanks for your significant comment. We have added some references in the revised version to broaden our literature survey, and the relative discusses are given as follows:

In [R.1], the authors focused on game-theoretical approaches in cognitive communication and networking systems, in which a stochastic anti-jamming game was proposed to design the optimal adaptive defense strategies against cognitive malicious attackers. In [R.2], the authors designed an opportunistic wireless communication setting and modelled the asset selling problem as a game theoretic variant in the completely observable and the partially observable cases respectively. In [R.3], the authors studied the problem of resource assigning in a single carrier communication system in the presence of a jammer, a Bayesian jamming game framework was proposed and the Nash strategy was also compared to the Stackelberg strategy to verify its sensibility. In [R.4], the authors studied a new anti-jamming problem of unknown nodes in a peer-to-peer network attacked by a random jammer or an intelligent one. In [R.5], considering the idealized case and potential energy constraint, the authors investigated the jamming attacks optimization problem that jammer could control the probability of jamming and transmission range to cause maximal damage to wireless network. Taking the perspective of Stackelberg game, the authors combed the anti-jamming technologies under different scenarios in [R.6], moreover, an anti-jamming decision-making framework was also proposed based on the adversarial characteristics between user and jammer, incomplete information and so on. The existed works investigated the anti-jamming problems from the perspectives of adaptive defense strategies, potential energy constraint, jamming modes, ideal and non-ideal cases in different communication scenarios to perfect the study. While game theoretical approaches, such as stochastic game and Stackelberg game, are just methods which can help us analyse the anti-jamming problems effectively. And the main difference is that, in the presence of an intelligent jammer, we investigate the relay selection problem under power energy constraint with imperfect information.

The related work discusses about the similar works done by other authors were given in the third paragraph in Section I. And we have pointed out some problems needing to be solved as follows: {However, considering the existing power optimization methods in anti-jamming domain, there are still some problems needing to be solved. Firstly, it is hard for user and jammer to get the math utilities and channel state information. Secondly, there exists error when user measures the communication quality or jammer measures the jamming effect. Thirdly, the optimizations of power strategy and relay selection need to be considered simultaneously}, which were the main differences between our work and other similar works in the fourth paragraph in Section I.

 

[R.1] Ray Liu, K. J.; Wang, B. Cognitive Radio Networking and Security: A Game-Theoretic View. Cambridge University Press. 2010.

[R.2] Naveen, K. P.; Altman, E.; Kumar, A. Competitive Selection of Ephemeral Relays in Wireless Networks. IEEE J. Sel. Areas Commun. 2017, 35, 586-600.

[R.3] Garnaev, A.; Trappe, W.; Petropulu, A. Combating Jamming in Wireless Networks: A Bayesian Game with Jammer’s Channel Uncertainty. ICASSP 2019.

[R.4] Garnaev, A.; Liu, Y.; Trappe, W. Anti-jamming Strategy Versus a Low-Power Jamming Attack When Intelligence of Adversary’s Attack Type is Unknown. IEEE Transactions on Signal and Information Processing over Networks. 2016, 2, 49-56.

[R.5] Li, M.; Koutsopoulos, I.; Poovendran, R. Optimal Jamming Attacks and Network Defense Policies in Wireless Sensor Networks. IEEE INFOCOM 2007.

[R.6] Jia, L.; Xu, Y.; Sun, Y.; Feng, S.; Anpalagan, A. Stackelberg game approaches for anti-jamming defence in wireless networks. IEEE Wireless Commun. Mag. 2018, 25, 120–128.

 

Point 3:

In the title, the words “Joint Optimization Approach” give a feeling that the user and the jammer selections are optimized jointly, however they are optimized hierarchically. The joint optimization is used for selecting a relay and a transmission power from the user side. Therefore, I prefer to change the title to “Hierarchical Optimization Approach”.

 

Response 3:

Thanks for your constructive comment. We have changed the title to “Hierarchical Optimization Approach”. In the anti-jamming relay communication system, the user and the jammer make strategy optimization hierarchically. Only for user, its relay and transmission power selections are optimized jointly. In our investigated anti-jamming problem, the central point is to realize the user’s strategy optimization with the change of jamming strategy. Considering the hierarchical competitive relationship between user and jammer, the title “Hierarchical Optimization Approach” is more accurate to describe the anti-jamming problem compared to “Joint Optimization Approach”.

 

Point 4:

As the symbols are defined in Table 1, there is no need to redefine them inside the text.

 

Response 4:

Thank you for your comments. In the revised version, we have simplified Table 1 and deleted some repetitive definitions.

 

Finally, we sincerely thank the reviewer-2’s positive comments and suggestions. We have tried our best to improve the manuscript and appreciate the reviewer-2’s warm work earnestly, and sincerely hope that the corrections will meet with approval.