FFANN Optimization by ABC for Controlling a 2nd Order SISO System’s Output with a Desired Settling Time
Round 1
Reviewer 1 Report
The work uses an interesting combination of models (FFANN and ABC) to improve the implementation of a control strategy for 2nd order systems (buck converter).
Ad FFANN (chap. 2/3)
The selected key parameterizations (number of neurons in hidden layer, activation function, weight optimization ) of the chosen FFANN model are not really explained or derived from the experimental setup
Ad ABC (chap. 4/5)
The (algorithmic) representation of the ABC optimization process for the control system under consideration (Fig.2!) is insufficient. In addition, the connection to Appendix B and to the general ABC description should be better documented/displayed.
Chap. 6
The results of the simulations with transfer functions are discussed very superficially. essentially, facts are presented. the quality of the control outputs with or without optimized FFANN parameters is documented solely by illustrations and qualitatively (is much better).
Chap. 8
Is too rudimentarily documented and can be deleted without loss for this paper.
Chap. 9/10
Insufficient and too superficial
In addition
the paper as a whole is written carelessly, not very reader-friendly and burdened with some inconsistencies (e.g. Fig 3 -> Fig 2, Fig 10 -> Fig 9, acknowledgement A should be integrated at line 189)
Author Response
Dear Reviewer 1,
Thank you for your suggestions about my paper.I have made the necessary arrangements within the framework of your suggestions. I think the article has a much stronger structure and academic knowledge with this modifications.
Thank you very much.
Aydin Muhurcu
Author Response File: 
Author Response.docx
Reviewer 2 Report
This manuscript proposed a feed forward artificial neural network to design the control strategy for 2nd order SISO system. To enhance the generalisation capacity of the proposed network, artificial been colony algorithm is employed to optimise the connection weights of the network. Finally, the Matlab Simulink platform simulation is used to validate the performance of the proposed control method. Overall, the topic of this work is interesting and the paper is well written. It can be considered to be published in Processes if the author can well address the following comments.
1. A comprehensive literature review on artificial neural network and swarm-based algorithms is required in Introduction.
2. I strongly suggest the author to cite the following paper, because it provides a good case on how the evolutionary algorithm can optimise artificial neural network.
Nonparametric modeling of magnetorheological elastomer base isolator based on artificial neural network optimized by ant colony algorithm, Journal of Intelligent Material Systems and Structures, 2015, vol. 26, no. 14, pp. 1789-1798.
3. There are a lot of optimisation algorithms that can be used to optimise ANN. The author should give the reason why the ABC is selected in this study. What is the superiority of ABC over other algorithms.
4. A flowchart is suggested to better illustrate the proposed ABC-FFANN method.
4. The weight results optimised by ABC may be locally optimal, which can result in the overfitting of the trained network. How can the author fix this problem?
5. A convergence analysis is required to analyse the performance of ABC.
6. More future work should be included in conclusion.
Author Response
Dear Reviewer 2,
Thank you for your suggestions about my paper.I have made the necessary arrangements within the framework of your suggestions. I think the article has a much stronger structure and academic knowledge with this modifications.
Thank you very much.
Aydin Muhurcu
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The author well answered the reviewer's comments. Hence, I suggest current version can be published in Processes.
