-
-\noindent\textcolor{black}{\textbf{MAJOR COMMENTS:}} \\
-
-\noindent {\bf 1.} Page 6, Section 3.2
-The author didn't explain how subregions are created. This is an important
-point, as clustering may have a significant impact on solution quality. Not only
-the size of the subregion should be discussed and analyzed, but also the
-clustering strategy.\\
-
-\textcolor{blue}{\textbf{\textsc{Answer:} In the study, we assume that
- the deployment of sensors is almost uniform over the region. So we only
- need to fix a regular division of the region into subregions to make the
- problem tractable. The subdivision is made such that the number of hops
- between any pairs of sensors inside a subregion is less than or equal
- to~3. In particular, we discuss the number of subregions in......}}\\
-
-
-\noindent {\bf 2.} Page 8
-The objective function (5) of the Mixed Integer Linear Program appears to be
-very questionable. Indeed overcoverage and undercoverage may compensate each
-other, so the same objective value may represent two incomparable situations. It
-seems that the semantic of the objective function is not well defined, as one
-may wonder what exactly is (quantitatively speaking) the problem objective.
-Coverage breach is obviously an issue in WSN, but why penalizing overcoverage? A
-two-phase approach where breach is minimized first, and then overcoverage is
-minimized would probably make more sense.
- \\
-
-\textcolor{blue}{\textbf{\textsc{Answer:} As mentioned in the paper the integer program is based on the model proposed by F. Pedraza and A. L. Medaglia and A. Garcia ("Efficient coverage algorithms for wireless sensor networks") with some modifications. Their initial approach consisted in first finding the maximum coverage obtainable from available sensors to then use this information as input to the problem of minimizing the overcoverage. But this two-steps approach is time consuming. The originality of the model is to solve both objectives in a parallel fashion. Nevertheless the weights $W_\theta$ and $W_U$ must be properly chosen so as to guarantee that the maximum number of points are covered during each round. By choosing $W_{U}$ very
-large compared to $W_{\theta}$, the coverage of a maximum of primary points is ensured. Then for a same number of covered points, solution with a minimal number of active sensors is preferred. }}\\
-
-
-
-\noindent {\bf 3.} Page 9
-In the MILP formulation, it is possible that some point p is never covered at
-all, which means that some part of the area to monitor may never be monitored by
-the WSN. The authors are referred to "alpha-coverage to extend network lifetime
-on wireless sensor networks", Optim. Lett. 7, No. 1, 157-172 (2013) by Gentilli
-et al to enforce a constraint on the minimum coverage of each point. \\
-
-\textcolor{blue}{\textbf{\textsc{Answer:} As previously explained, the model with the appropriate weights ensures that a maximum number of points are covered for the set of still alive sensors. The coverage is measured through the performance metrics "coverage ratio". The coverage ratio remains around 100\% as long as possible (as long as there are enough alive sensors to cover all primary points) and then decreases. The problem introduced in "alpha-coverage to extend network lifetime
-on wireless sensor networks" by Gentilli is quite different. In this problem, the coverage ratio is fixed to a predetermined value ($\alpha$) and the amount of time during which the network can satisfy a target coverage greater than $\alpha$ is maximized. }}\\
-
-
-\noindent {\bf 4.} Page 13
-The criterion "Energy Consumption" is the average consumption per round. But the
-duration of a round is a feature that can be arbitrarily set in the algorithm.
-Computing the average energy consumption per unit of time over the network
-lifetime would be better, as it is independent from the number and duration of
-rounds. \\
-
-\textcolor{blue}{\textbf{\textsc{Answer :} Yes, you are right. It is possible to obtain the average energy consumption per unit of time by dividing the criterion defined in section .... by the round duration.}}
-
-
-
-\noindent {\bf 5.} Page 15-18
-Figures 2-6 mention four different versions of MuDiLCO. The performance of these
-different versions should be analyzed with more details for each figure.
-Alternatively, the authors may remove some versions of MuDiLCO if they do not
-bring any valuable insight. \\
-
-
-\textcolor{blue}{\textbf{\textsc{Answer :} Right. We have completely re-written section 5 to highlight the most significant results. }}
-
+\noindent\textcolor{black}{\textbf{MAJOR COMMENTS:}}\\
+
+\noindent {\bf 1.} Page 6, Section 3.2 The author didn't explain how subregions
+are created. This is an important point, as clustering may have a significant
+impact on solution quality. Not only the size of the subregion should be
+discussed and analyzed, but also the clustering strategy.\\
+
+\textcolor{blue}{\textbf{\textsc{Answer:} In our work, we assume that the
+ sensors are deployed almost uniformly and with high density over the region.
+ So we only need to fix a regular division of the region into subregions to
+ make the problem tractable. The subdivision is made such that the number of
+ hops between any pairs of sensors inside a subregion is less than or equal
+ to~3.
+% In particular, we discuss the number of subregions in......
+}}\\
+
+\noindent {\bf 2.} Page 8 The objective function (5) of the Mixed Integer Linear
+Program appears to be very questionable. Indeed overcoverage and undercoverage
+may compensate each other, so the same objective value may represent two
+incomparable situations. It seems that the semantic of the objective function is
+not well defined, as one may wonder what exactly is (quantitatively speaking)
+the problem objective. Coverage breach is obviously an issue in WSN, but why
+penalizing overcoverage? A two-phase approach where breach is minimized first,
+and then overcoverage is minimized would probably make more sense.\\
+
+\textcolor{blue}{\textbf{\textsc{Answer:} As mentioned in the paper the integer
+ program is based on the model proposed by F. Pedraza, A. L. Medaglia, and
+ A. Garcia (``Efficient coverage algorithms for wireless sensor networks'')
+ with some modifications. Their initial approach consisted in first finding
+ the maximum coverage obtainable using the available sensors and then to use
+ this information as input to the problem of minimizing the overcoverage. But
+ this two-steps approach is time consuming. The originality of the model is
+ to solve both objectives in a parallel fashion. Nevertheless the weights
+ $W_\theta$ and $W_U$ must be properly chosen so as to guarantee that the
+ maximum number of points which are covered during each round is maximum. By
+ choosing $W_{U}$ very large compared to $W_{\theta}$, the coverage of a
+ maximum of primary points is ensured. Then for a same number of covered
+ primary points, the solution with a minimal number of active sensors is
+ preferred. }}\\
+
+\noindent {\bf 3.} Page 9 In the MILP formulation, it is possible that some
+point p is never covered at all, which means that some part of the area to
+monitor may never be monitored by the WSN. The authors are referred to
+``$\alpha$-coverage to extend network lifetime on wireless sensor networks'',
+Optim. Lett. 7, No. 1, 157-172 (2013) by Gentilli et al to enforce a constraint
+on the minimum coverage of each point.\\
+
+\textcolor{blue}{\textbf{\textsc{Answer:} As previously explained, the model
+ with the appropriate weights ensures that a maximum number of points are
+ covered by the set of still alive sensors. The coverage is measured through
+ the performance metrics ``coverage ratio''. This one remains around 100\% as
+ long as possible (as long as there are enough alive sensors to cover all
+ primary points) and then decreases. The problem introduced in
+ ``$alpha$-coverage to extend network lifetime on wireless sensor networks''
+ by Gentilli is quite different. In this problem, the coverage ratio is fixed
+ to a predetermined value ($\alpha$) and the amount of time during which the
+ network can satisfy a target coverage greater than $\alpha$ is
+ maximized.}}\\
+
+\noindent {\bf 4.} Page 13 The criterion ``Energy Consumption'' is the average
+consumption per round. But the duration of a round is a feature that can be
+arbitrarily set in the algorithm. Computing the average energy consumption per
+unit of time over the network lifetime would be better, as it is independent
+from the number and duration of rounds.\\
+
+\textcolor{blue}{\textbf{\textsc{Answer :} Yes, you are right. It is possible to
+ obtain the average energy consumption per unit of time by dividing the
+ criterion defined in section~4.3 by the round duration.}}
+
+\noindent {\bf 5.} Page 15-18 Figures 2-6 mention four different versions of
+MuDiLCO. The performance of these different versions should be analyzed with
+more details for each figure. Alternatively, the authors may remove some
+versions of MuDiLCO if they do not bring any valuable insight. \\
+
+\textcolor{blue}{\textbf{\textsc{Answer :} Right. Therefore we have completely
+ re-written section 5 to highlight the most significant results.}}
