Joint Optimization of Task Caching and Computation Offloading for Multiuser Multitasking in Mobile Edge Computing

Round 1

Reviewer 1 Report

The text is presented with scientific rigor and practically all stages of the proposal are presented clearly and well illustrated.

  As a way of contributing to the work, here are some small suggestions to consider:

- especially in the initial part of the text, the authors use very long sentences.

- DQN on line 115 (explain meaning on first use). Same situation for DDQN.

- The use of the term "Literature" and the reference number in the text is not pleasant for the reader.

- what are the criteria for defining the parameters presented in table 2?

- better explain the example used in the tests (cameras). What tasks were performed (which were repeatedly used in testing?). What real situations do they reproduce?

- present the computational time complexity of the algorithm.

Author Response

Dear Reviewer.
Thank you for taking time out of your busy schedule to review my manuscript and for your professional comments and suggestions. Your feedback was very valuable and enabled me to think more deeply and improve my work.

I have carefully considered each of your comments and made corrections. My response to the review comments has been uploaded as an attachment.

Thank you again for your support and help!

Author Response File: Author Response.pdf

Reviewer 2 Report

In this work, the authors address the problem of caching and offloading on mobile edge computing in an effort to reduce latency and energy consumption in order to meet the user’s needs. To achieve this, they propose an optimized version of deep reinforcement learning (P-DDPG) algorithm.

To sum up, the problem is practical and helpful, and the authors describe and present an optimization model, and assess its performance. It's a well-written and cohesive work that readers may find helpful. Also, the authors support their work with experimental measurements and quantify the efficiency of the proposed solution.

In my opinion, it is very good work that can be published in this form.

Author Response

Dear Reviewer:

Thank you very much you for reviewing my paper and giving your valuable comments. Your favorable comments are a great encouragement to me and a motivation for me to continue my efforts. My response to your review comments has been uploaded as an attachment.

Finally, thank you again for your recognition and guidance, and I will continue to work hard to improve the quality of my research and academic level.

Thank you again for your support and help!

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper proposes and assesses a deep reinforcement learning-based algorithm to take decisions about task computation and caching that aims to minimize the delay and energy consumption of the task processing process.

The paper clearly states the considered scenario and problem. The authors exhaustively survey the main reference papers for the literature stating what the present contribution offers more. The considered system and the proposed algorithm are well described in detail. The performance evaluation shows the obtainable performance in terms of different metrics and in different scenario conditions and highlights the achievable performance improvement compared to the other three solutions from the state-of-the-art.

The paper deserves to be published, after addressing the following few minor concerns:

1.     In Section 5, figures 4, 6, 7, and 8 show the obtained performance in terms of the “Average System Cost (J)”. However, this cost has not been clearly defined in any of the previous sections. Is it the average cost defined in Eq. (11) over multiple time slots? Or is it something else?

2.     In Section 5, add some details about the other three considered strategies (PCL, RCAO, DDQN) to let the readers know how they operate from a high-level viewpoint. Adding one reference per strategy would also help.

3.     In Section 3.1, the number of tasks should be K instead of k, as the number of users is M.

4.     In Figure 1, the index used to identify the tasks and the users should be changed in i and j, respectively, in accordance with the letters used throughout the text.

5.     In Figure 5, there is a typo to fix in the y-axis label: “Delary” -> “Delay”

Author Response

Dear reviewer.
I would like to express my sincere gratitude to you for your careful review of my thesis and your valuable comments. Your professional advice has played a key role in improving the quality of my thesis.

I have carefully revised the paper in the light of your comments and have made comprehensive adjustments and improvements to the manuscript. I am convinced that these revisions have not only improved the quality of the paper, but also better responded to the expectations of the reviewers. My response to your review comments has been uploaded as an attachment.

Thank you again for your support and assistance!

Author Response File: Author Response.pdf