Balancer Genetic Algorithm—A Novel Task Scheduling Optimization Approach in Cloud Computing

Round 1

Reviewer 1 Report

1. The first paragraph contains much common sense for Cloud services. Reduce or summarize for key point,
2. Is “TSGA” term same as “TS-GA” in page 4?
3. The term “PSO”, I guess, is “Particle swam optimization”, isn’t it? Then, use “PSO” with “Particle swam optimization (PSO)” at first appearance.
4. There are so many mis-typing in this paper. Correct all.
5. Your paper focuses on only GA variational algorithm for load balancer. Isn’t there any paper in which a neural network technique is applied for load balancing of Cloud? Add related works with deep neural network in your “related work” section.
6. The parameter defined for your algorithm 1 is not necessary because it depends on a target. So, it should be defined in the experimental section with test data.
7. In algorithm, s<-1 (line# 18) is overwritten with s <-(SizePopulation/2) -2 (line # 24). Is this correct?
8. I don’t get the algorithm 1 because there is no explanation for sub-algorithm such as SelectAlgorithm(load_balancer[]).
9. Also, I don’t know why you used constants like 0.2 and/or 0.8 in the algorithm. If possible, use a symbol instead of a constant. The constant is selected at the time of experiment.
10. Is the equation (1) correct? Is it ? Is the unit of makespan same AVG_psu?
11. Is there no weight parameter for AVG to make balance between make and and AVG?
12. What is the unit of PSU_j? Is it between 0 and 1 so that  ?
13. What make the PSU in equation (3)? Does it make to percent unit?
14. In equation 12, if PSU_j exceed 200, what is happening?
15. What is the notation VM_j.mips in the equation (7)
16. Divide the paragraph of selection operator (middle of page 9) to a new section number. It cannot be included fitness section.
17. What is “s” in “To compute sum s” sentence of page 9. Define “s” first what it is.
18. What is full terminology of “ARUR” ? Is it used at any papers for performance evaluation ? And also, if you want to use this measure, you should specify this measure at section 4 with “makespan”.
19. Use another paragraph or section for “ARUR” experimental result.
20. Did you analyze time complexity of your algorithm with heuristic algorithms? I guess that your algorithm uses high complexity because GA algorithm requires multiple data(chromosomes).
21. For overall, your paper needs to correct words, to reconstruct paragraph and equations. And, also, need to refer neural network technique which is used at load balance.

Author Response

We would like to thank the editor and the reviewer for their time and for providing us extremely useful comments/ feedback. We have thoroughly revised the manuscript as per the suggestions of the reviewer. We hope that the editor will find the paper suitable for publication.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper studies the task scheduling problem in cloud computing. A Balancer Genetic Algorithm (BGA) is proposed to solve it. Overall, it is good research topic. The writing structure is well-organized. Some changes are required to further improve it. 1. P2, please provide the full name of MGGS at its first appearance. 2. P4, please provide the full name of LCFP and SCFP. 3. P4, please provide the full name of LAGS at its first appearance. 4. The following express should be improved.” RMGO a genetic based load balancer was proposed in [14],…” Furthermore, what is RMGO? 5. Please provide the full name of AIGA. 6. A heuristic algorithm is utilized in this study. However, the reason of proposing a new one is not sufficiently provided. Literature review on existing heuristic algorithms, e.g., GA, Particle swam optimization (PSO), Ant Colony Optimization (ACO) etc., is not sufficient. Some references should be noted. (1) Comprehensive learning particle swarm optimization algorithm with local search for multimodal functions. IEEE Transactions on Evolutionary Computation 23 (4), 718-731, 2019. (2). An Effective Cloud Workflow Scheduling Approach Combining PSO and Idle Time Slot-Aware Rules. IEEE/CAA Journal of Automatica Sinica, 8(5), 1079-1094, 2021. (3) Enhanced particle swarm optimization for task scheduling in cloud computing environments. Procedia Computer Science, 65, 920-929, 2015, etc. 7. Please improve Section 1 and Section 2 to concentrate on the motivation and background of the studied problem and utilized approach. Some irrelated information can be removed. 8. “Modified Genetic Algorithm (MGGS) [12],” Is it correct to use MGGS for this? Where is “s”? 9. “In [21] SJFP a heuristic is merged in PSO…” What is SJFP? 10. P7, line numbers in Algorithm 1 should be corrected. 11. P8, the format of Table 1 should be adjusted. 12. “BGA is a multi-objective technique in which two major goals are addressed, one is makespan and other is load balancing.” Please check the grammar in this sentence. 13. P10, the name of Algorithm 2 should be improved to be more accurate. 14. P15, Table 2 and 3, the title should before the table. 15. There are many abbreviations in the whole text. Even they are from literature, please provide their full name. This needs to be checked thoroughly. 16. Figure 2, it is not proper to add a box for next generation. Please connect the out part to the exact process in this part. 17. Figure 3 can be removed. This process can be expressed clearly in text.

Author Response

We would like to thank the editor and the reviewer for their time and providing us extremely useful comments/ feedback. We have thoroughly revised the manuscript as per the suggestions of the reviewer. We hope that the editor will find the paper suitable for publication.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

See an PDF attached file.

Comments for author File: Comments.pdf

Author Response

Please find the replies in the attachment.

Author Response File: Author Response.pdf