Evolutionary Optimization of Drone-Swarm Deployment for Wireless Coverage

Round 1

Reviewer 1 Report

This paper studied the drone swarms deployment problem to minimize the maximum energy consumption among all drones while achieving full coverage over a target area. The authors proposed a new approach of using an integer code scheme to encode the sequence of deployment of the drones. The energy consumption incurred during the horizontal and vertical flying time is adopted as the fitness value and a feasibility checking algorithm is introduced. Apart from this, a multiobjective optimization method is proposed to find the tradeoff among three objectives. The paper is well-written and easy to follow. It is a extended version of a preliminary conference paper by adding substantial new material. This research could be serve as a foundation for research involving the optimization of wireless coverage. There are still some minor issues needed to be addressed. 

1.There are some typos and presentation issues needed to be resolved. Maybe some processional reading-proof is required.

2.The Fig 1 should be improved by revising the fonts and improving illustration quality.

3.The equation (2) should be revised and corrected.

4.The drones or UAVs should be consistent through the whole manuscript. Please check.

Author Response

Thank you very much for your review. Please find the attachment for the details.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper minmizes the maximum energy consumption in the considered drone swarms deployment problem, which aims to balance the energy consumption of all drones and maximize the full coverage network lifetime. A genetic algorithm is present to search a favorable solution. Moreover,  the tradeoffs between energy consumption, number of drones, and coverage rate of the target area are investigated. A drone swarms deployment algorithm based on MOEA/D is proposed to find the best tradeoff between these objectives. The work is solid and meaningful. However, there are some minor comments as follows:

In the abstract, what is MOEA/D? please give the full name before the abbreviation. 

In the introduction, the sentence "Since the wireless coverage of the drone is intractable" (line 37) is unclear, it should be "Since drones' wireless coverage problem is always intractable" . Also, in line 46, "drones’s deployment" should be "drones' deployment".

 

 

 

Author Response

Thanks a lot for your valuable comments. Please find the attachment for the answers.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper describes an algorithm for deploying drones efficiently over a target area. In particular, finding the way to the target point of the drone is discussed, as the authors believe that this consumes a large part of the available energy. By optimizing the route, the time the drones spend in the target area should be maximized.
The authors are of the opinion that the DSDP has not been adequately addressed in previous publications and that this is a novelty. Although the DSDP is a classic optimization problem, as the authors also state, it is understandable why other disciplines dealing with efficient path selection have not considered it. Another look would have been desirable here.
The proprietary algorithm for deploying the drones is then presented. I have to admit that I couldn't follow the description. The presentation of the algorithm in Figure 3 is not easy to understand. It is also not clear what vi and ci from Equation 2 are. Unfortunately this continues. In figure 8 the crossover of two individuals is shown, but in the end PB' has more gray boxes than before. For me, parts of the explanation are missing here.
The evaluation is also not understandable for me. Does the runtime matter? I assume, the routes are calculated beforehand? 850 secs runtime is really a problem? The subproblems are also not introduced. It seems to be necessary to know NSGAII, SPEDII or MOED/D better. Furthermore, It is not clear what is shown in Figure 13 or 14. The form of presentation is not appropriate for me.
The paper may present an interesting aspect of drone communications. Unfortunately, the algorithm is not comprehensible for me as a reader. This may be due to a lack of knowledge in the subject area. However, the presentation should be comprehensible even without special knowledge of the subject. The presentation should be improved here.

Author Response

Please find the attached PDF document for the point-by-point response. Thanks for your review and hard work.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Something still seems to be wrong in section 2.2.

Line 120 says a solution instance is encoded as a sequence? What is that supposed to be? One solution is the set of positions of the drones, this was defined as di(yi, hi). Does each I1...In now stand for such a drone position?

What is N~? Where do the solution sets come from?

What shall the ith parent be? Then what should Ia be? If this is an element from Sequence I or if this is such a sequence, then the naming is very unfortunate. What is the other parent I~?

What is a gene now? If this is such a sequence, then lines 125 to 130 make sense, but then line 133 makes no sense, there only two elements of a sequence are exchanged. However, line 133 says that two nonidentical genes are swapped. Why nonidentical, is there a possibility in a sequence that two elements are the same? But that would mean that two drones are in the same position.

Why is the 5 not included in Ib now? I think there is some kind of typo here. Otherwise I don't understand the offspring.

For me, Formula 2 has the problem that ri has to be dependent on hi. If not, then I get the greatest distance when hi = 0. That seems illogical to me. Where is he from?

I still think that the identifiers are very unfortunate. So in algorithm idx is the index. But doesn't idx have to have an energy budget if idx results from high and high is initially assigned Emax? In Algorithm, Ib is element of Di. What is Di? Is this the initial solution set? Where does it come from?

Unfortunately, a great many of these passages are in the paper. Minor improvements have been made compared to the last version. However, many parameters are still not correctly defined, so that the formulas and algorithms are extremely difficult for reader to understand.

Author Response

Thank you very much for your hard work and valuable comments. We have revised the paper accordingly. Please find the attachment for the responses.

Author Response File: Author Response.pdf