MiA-CODER: A Multi-Intelligent Agent-Enabled Reinforcement Learning for Accurate Coverage Hole Detection and Recovery in Unequal Cluster-Tree-Based QoSensing WSN

Round 1

Reviewer 1 Report

In this paper, the problem of coverage hole in wireless sensor networks is studied in detail and the solution is provided. Firstly, the energy optimization is used to prevent the formation of coverage holes, then detect coverage holes by using Virtual Sector based Hole Detection (ViSHD) protocol, and the reinforcement learning algorithm is designed to repair coverage holes. A large number of experimental results are presented to show the effectiveness of the proposed method. However, this paper is not well written. Sometime it is hard to follow the authors' idea. So I suggest the paper should be modified carefully before publishing.
Some suggestions are as follows,
1.Since there are several aspects to consider in the paper, they should be clearly described, such as the expression of various notations and their meanings, it is suggested to use notations table.
2.The expression of the formula can be improved to express more clearly.
3.Figure 6-9 are recommended to be redrawn to ensure good appearance.
4.The contents of some figures and tables in the paper are repeated, as shown in Figure 6 and Table 5.
5.There are several algorithms in this paper, and their complexity can be analyzed.
6.Pay attention to the consistency of the whole text, such as Fig.x and Figure x, Table 3 and Table III.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Title: MiA-CODER: A Multi-Intelligent Agent-Enabled Reinforcement Learning For Accurate Coverage Hole Detection and Recovery in Unequal Cluster-Tree-Based QoSensing WSN 

Comment: Some of the minor queries are:

1. This is not clear how the author derives Eq 1. (remaining energy of the sensor node is expressed as..)
2. Explain: "The remaining energy for all the nodes is calculated to elect the optimal cluster head."
3. Eq 14-18 are derived by authors or cited from existing literature?
4. Discuss the computation complexity associated with the simulations.
5. Are there any limitations on the number of static nodes and the number of mobile nodes? explain.
6. What is the minimum energy assumed during initial round simulation?
7.  the authors may include recent works related with multi-agent models implemented in WSN. ex: "Multi-agent-based smart power management for remote health monitoring"; "Energy-Efficient Dynamic Clustering for IoT Applications: A Neural Network Approach"; "On-demand efficient clustering for next generation IoT applications: A Hybrid NN approach"; etc. These references are suggested based on their relevancy with the proposed energy efficient models and clustering in WSN.
8. In Fig. 6., the simulation starts after 30s. The reason is not not clear. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Please to see the attached file. Thank you.

Comments for author File: Comments.pdf

Author Response

Please see the attchment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The revised manuscript of the paper has basically solved the problems raised earlier and made a reasonable explanation for the problems, but there are still some small problems in the format of the paper that have not been modified. It is suggested to make minor revision to improve the quality of the paper. Here are the problems that need to be improved:
1. The symbolic expression of the formula, for example, the CH-intra (en) of formula (5) may be regarded as a minus;
2. Reference format of figures, such as "Fig.3" on page 14, "figure 5" on page 18 and "figure .10" on page 22.

Author Response

Please see the attachment

Author Response File: Author Response.docx