A Path Planning Model with a Genetic Algorithm for Stock Inventory Using a Swarm of Drones

Round 1

Reviewer 1 Report

Please read the attachment. Thank you.

Comments for author File: Comments.pdf

Author Response

Answers to the reviewer's opinion

First of all, thank you very much for the very thorough and constructive review. We received many useful ideas and suggestions.

1. Thank you, we fixed it.
2. Thank you, we fixed it.
3. Thank you, we fixed it.
4. In this subsection, the route and processing time of the two compartments to be processed one after the other, which are located in different rows, are given.
5. Condition 1 and condition 2

Condition 3

Condition 3

6. Added: In certain warehouses - due to safe traffic - the corridor between individual shelf systems can only be walked in one direction. Generally, the direction of travel in two adjacent corridors is opposite, so at the end of the corridor you can come back in the corridor immediately next to it. Based on these, the direction of traffic is the same in every other corridor, so we can say that the direction of travel is opposite in even-numbered and odd-numbered corridors.

Several warehouses have one-way traffic. The direction of travel is opposite in odd-numbered and even-numbered aisles.

7. Thank you, we have fixed:

Condition 5 and condition 6

Aisles change in case of one-way traffic 5.-6. conditions

8. We fixed it.
9. Thank you, we fixed it.
10. Corrected.
11. Thank you. We provided an explanation:

Note. In the following, we will use the “:=” operation. This has a different meaning than “=”. “:=” means that the value of the variable on the left side of the operation is equal to the expression on the left side.

12. Thank you, we fixed it
13. Thank you, we fixed it.
14. Thank you, we fixed it.
15. Thank you, we have changed it
16. Thank you, we have amended and added the following:
17. Conclusion

In this article, we presented the mathematical model and a solution method for inventorying a multi-user warehouse with drones.

Our tests and experiences so far have shown that the given mathematical model describes well the majority of warehouses in practice and the task of inventorying with drones. This also shows that in practice the model can also be used in logistics decisions.

As described in the previous chapter, we experienced certain problems during the solution method, which fortunately did not cause problems in the case of the warehouses we examined, however, in order to ensure the stability of the solution, further tests must be carried out.

Questions

Question 1.

During the mathematical modeling, we followed the principle of top-down progress to the detailed model. For this we used different model levels.

In solving the tasks related to the model, we examined several methods:

1. traditional programming methods (within the field of operations research). We tried to simplify the model in such a way that we get an integer programming task. However, reducing the objective function did not bring really good results, so we rejected this solution.

The 2nd swarm intelligence method did not really lead to results, but at the same time it gave a very good starting value, which can help the GA solution.

3. We also prepared the Flower Pollination Algorithm program, but this is still under investigation, in terms of efficiency and results.
4. We have also investigated risky solutions, with varying degrees of success.

Question 2.

We have created a diagram of the most important algorithms. The diagrams of the other algorithms mostly contain sequences.

Question 3.

Yes. We also had this problem. This is why the one-corridor-one-drone condition was included. Another critical point is the corridor change. We solved this so that the drones start their work at different heights, i.e. they have different height channels. It starts with an ascent at the start. When changing lanes, each drone can only travel in its own channel.

Question 4.

Only partially. There were no significant problems in the examined warehouses, but they may occur. Because of this, the channels of the drones. It is determined based on security considerations. Movement along the center line of the corridors is also aimed at reducing distractions. A software will continuously monitor their operation.

Question 5.

The measurement results and the examination of the efficiency are beyond the scope of the study, since we wanted to outline the model and the solution method. At the same time, we have such results. We would like to present these after several checks. It can be seen that the traditional inventory involves limiting the warehouse traffic, which can also be done continuously at night. Also, the inventory can be done in at least a third of the time. It used to be done by several people in one warehouse, with this method it can also be solved with one dispatcher.

Question 6.

The run was performed in MatLab in the first round. The final application will be done by developers. The routines we wrote are available, but they are quite large and specially designed for MatLab. At the moment I don't know how to attach them.

Question 7.

We shortened the Conclusion. The work was shared.

Author Contributions: Conceptualization, M.G.; methodology, M.G.; writing—review and editing, J.U.; visualization, J.U.; investigation, M.G. All authors have read and agreed to the published version of the manuscript.

 

Author Response File: Author Response.docx

Reviewer 2 Report

This manuscript presents a detailed introduction section, but authors must clearly state the contribution of this paper. I think, the presentation of contributions of this paper is weak.

The manuscript includes a lot of equations. Authors must give the explain of parameters of equations in the end of equations or they must add to a nomenclature table end of the manuscript.

A flow chart that defines the methodology proposed this paper can be add in material and methods section. Especially, this flow chart must include three step of the problem that defined by authors.

There are not explain of parameters in figures and equations. So, some part of your paper can be not understood clearly.

In figure 10, the colour of obstacle and drone must be different.

What are the advantages of the new model you propose over the ones in the literature?

How do you do validation of this new model?

Why did authors use genetic algorithm for this problem? What is the advantage of genetic algorithm according to other new meta-heuristics algorithms?

What is the difference between your study and this study (‘A Path Planning Model for Stock Inventory Using a Drone’).

I did not understand where use GA in this problem.

Some figures in the manuscript are never mentioned in the text.

 

What is meaning the objective function of this study?

 

 

Author Response

Dear Reviewer!
Thank you for the useful comments!

(1)This manuscript presents a detailed introduction section, but authors must clearly state the contribution of this paper. I think, the presentation of contributions of this paper is weak.

Thank You. We made the following addition:
In this article, we would like to present a mathematical model and procedure that we have not encountered during the examination of the literature. The model and method show the inventory of a high-rise warehouse that cannot be reached by GPS with multiple drones using a QR code, which has multiple users. The complexity of the problem will be presented later.

(2) The manuscript includes a lot of equations. Authors must give the explanation of parameters of equations in the end of equations or they must add to a nomenclature table at the end of the manuscript.

The nomenclature tables can be found summarized at the beginning of the article with detailed explanations. (See Table 1. Table 2. Table 3.). They explain all the parameters in the equations together with the indices. We intended to present the explanation of the parameters at the beginning of the article, since this way we can refer to the individual parameters later on.

(3) A flow chart that defines the methodology proposed in this paper can be added in the material and methods section. Especially, this flow chart must include three steps of the problem defined by authors.

The flowchart was prepared for the main process.

(4) There are no explanations of parameters in figures and equations. So, some part of your paper may not be clearly understood.

Based on the previous comment, all parameters are explained in the summary tables (Ld Table 1. Table 2. Table 3.). in tables. As usual in mathematics, we first give the basic concepts, then the definitions based on this (see table), and then we derive the relationships from this. Therefore, in our opinion, repeating the definition is no longer justified here.

(5) In figure 10, the color of obstacle and drone must be different.

Thanks, fixed.

(6) What are the advantages of the new model you propose over the ones in the literature?

The advantage of the method is that in mixed high-rise warehouses, where there is no GPS availability and the goods of different users are mixed, the inventory can be made automatic and fast with the help of drones. As a result, the traditional storage functions of warehouses only have to be interrupted for a short time. We have not come across such a solution in the literature.

(7) How do you do validation of this new model?

Validation is in progress for several specific warehouses. In practice, we examine several warehouses. This process is currently ongoing.

(8) Why did authors use genetic algorithm for this problem? What is the advantage of genetic algorithm according to other new meta-heuristics algorithms?

Previously, we examined several solution methods - as described - but some of them did not give a stable solution, while others gave a local optimal solution during the search. During our tests, GA solved the problem quickly and well, and the optimal solution could be put back into practice.

(9) What is the difference between your study and this study (‘A Path Planning Model for Stock Inventory Using a Drone’).

The significant difference is that the previous study can manage a warehouse with a fixed structure. Shelf systems with variable compartment sizes are not. Many warehouses are like this, but the new problem arose when we encountered a specific warehouse where the shelves were different even in adjacent columns and several shelf sizes appeared even within one column. The previous solution cannot be used here. They also planned to use more drones in this warehouse, which fundamentally changed Policy's mathematical model of crawling routes. The objective function has also changed. Previously, it was a simple time function, now it was necessary to minimize the maximum of several functions. In addition, access has also changed due to the fixed routes. These are described in the references of article [23].

(10) I did not understand where to use GA in this problem.

The GA is used to minimize the objective function (i.e. the time function), otherwise, to determine the optimal traversal paths of each drone according to time. (in the unamended article, it is described in lines 434-445). We also created a separate software for this (which is not part of the study). Subchapter 3.9 specifies based on the model. All of the specified auxiliary functions check the appropriate limiting conditions of the GA (3.9.2) and calculate the fitness function (3.9.1). Chromosomes are route assignments per drone. We worked with a population of different sizes, we also examined running results and times. This is not part of the study.

(11) Some figures in the manuscript are never mentioned in the text.

Thank you, we will provide:
The following figures (Figure 3, Figure 4, Figure 5) show the main parameters of the structure of the warehouse to be modeled.

Let x denote the horizontally mapped path of the drone's path, y the vertically mapped path of the drone's path (Figure 6)
The above periods are detailed in this chapter (Figure 7).
3.6.1. The distance and time from the docking station to the starting point (Figure 8)
The time for photography is added at the end (Figure 9).
3.6.4. From one compartment of one row to another compartment of another row (Figure 10):
The structure of the specific warehouse (Figure 11, Figure 12):

(12) What is the meaning of the objective function of this study?

Thank You. The objective function represents the duration of each drone's visit to the allocated compartments and the charger, and to take photos. According to this article, it depends on the access route. According to the article, the duration of the entire inventory is the maximum of the inventory durations of the individual drones. This should be minimized. This was described in detail in relations (6)-(13). For the sake of better transparency, we insert the following sentence under (13):
This is the objective function of the mathematical model.

Author Response File: Author Response.docx

Reviewer 3 Report

The article is presenting interesting idea but the authors should add a section with simulations to verify their approach. 

Author Response

Thank you for your comment. we fully agree with the comment. It has been explained in detail to the other reviewer that the solution method and the effectiveness of the solution are still being tested. The runs so far have confirmed our expectations, but until validation is done (due to possible biases and errors) we do not wish to present it. In this paper, we wanted to present the model and the solution, and to show that the problem can be addressed with this approach.

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear Authors and Editor:

Thank you for your response. 

The authors have answered and corrected all my questions and comments. 

One more point is under consideration. You need to revise your references regarding algorithm applications. Please consider adding these topics to your literature review for your method ( for example, CFD Analysis and Optimum Design for a Centrifugal Pump Using an Effectively Artificial Intelligent Algorithm; Optimizing compliant gripper mechanism design by employing an effective bi-algorithm: fuzzy logic and ANFIS).

Thank you for reading. 

Best regard; 

The reviewer.

Author Response

Thank you for your review and we have considered your suggestion and incorporated it into our literature research.

Author Response File: Author Response.docx

Reviewer 2 Report

Thanks for your reply.  I still think that the contribution of this paper is not emphasized enough in the introduction section. I also recommend that you do a detailed reading in terms of grammar and readability.

Author Response

Thank you for your review, it had been corrected. It had been checked by a native English teacher.

Author Response File: Author Response.docx

Reviewer 3 Report

I accept author's explanations. 

Author Response

Thank you for your review.

Author Response File: Author Response.docx