Two-Level Programming Model Based on Cooperative Operation Study of Stakeholders in Hazardous Chemical Storage

Round 1

Reviewer 1 Report (Previous Reviewer 1)

All of the comments are well addressed. The paper is ready for publication.

Author Response

Thank you for your review!

Reviewer 2 Report (New Reviewer)

There is no doubt about the importance of hazardous chemical storage and management. To ensure storage safety, and balance the interests of stakeholders, the authors propose a two-level programming model considering the maximization of social welfare and the interests of warehousing enterprises. The upper model mainly refers to the regulatory department represented by the government with cost related coefficient/parameters. In the lower model, the comprehensive risk level of the warehouse is determined by the warehouse enterprise. The relevant case analysis and comparisons show some good results and insights presented in this paper. The topic is of importance. The model is good presented. The results are given and supported by case and comparisons. The literature part should be updated, better organized.

Details are as follows:

(1)Literature review. There are many literature presented in ths paper, but it should be better classified and updated;

(2) table 1 looks not good and should be rewritten;

(3)the authors present some case of Shanghai Beifang Storage and Transportation Group. The results still be valid for other company? why?

(4) the Conclusion, it should be written to be in contract to the previous model and analysis;

(5)how to obtain the parameters(such as those in Table 2-7)?

(6)it is suggested to updated the literature.

Author Response

Plaese see the attachment. And we sincerely accepted the reviewer's suggestions on the English expression of our paper. We also invited Associate Professor Yongping Zhang from California State Polytechnic University, Pomona, to revies the English expression and logic of the papaer. As shown in the revised section of the paper.

Reviewer 3 Report (New Reviewer)

The novel two level model formulated by the authors helps to understand the cooperative operation among several stakeholders in hazardous chemical industry. The idea to minimize the overall operational costs of the warehouse enterprise within the boundaries of social costs minimized through Government policy is an interesting approach and this approach may help the enterprises to justify their operational costs including supervision and punishment costs.

The formulation of the model is neatly described. The quantitative values of risk impact and risk level are based on ordinal scale and while this works for the model, the researchers in future studies can explore using better methods to represent the underlying risk. Also I feel the iterative performance of the algorithm would have been presented and discussed to give more clarity to the readers.

While the authors claim that the results re from a single case, I feel the model can be easily extended to similar situations and the authors would provide some insights on how to tweak the model parameters to represent some more generic cases. This would greatly enhance the generalizability of the model. The authors would have also described a little more about how the various coefficients can be used by practitioners for managerial decision making. Possibly the manuscript word limit prevents the authors to do so.

The overall logic and presentation of the manuscript is acceptable.    

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Hazardous chemical storage, and their interest appeals of relative stakeholders were discussed in this paper, and a two-level programming model considering the maximization of social welfare and the interests of warehousing enterprises was proposed, which can also balance the interests of stakeholders and achieve cooperative operation. It’s an interesting topic, the structure of paper seems good, and the conclusion seems also reliable. Some comments in my opinion may be useful to improve the paper:

1.     How to determine the risk level like a1,a2, a3 in formula(3)? Some explanation should be given.

2.     About the interest of different parts, game theory is a good method, difference between the method in this paper and game theory should be given.

3.     What is the limitation of the proposed method? Please provide some comments.

4.     English writing should be improved for easy reading, for example, the following sentence in abstract “in the lower level, the warehousing enterprise first judged the comprehensive risk level of the warehouse, and then determined the supervision coefficient, and through the supervision coefficient and punishment coefficient, determined the warehousing operation cost, warehousing supervision cost and punishment cost when the accident occurs under different risk levels”, some subordinate clause are suggested to replace a so long sentence.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors observed an interesting topic of a model for cooperative operation of stakeholders in hazardous chemical warehousing. However, the submission has significant drawbacks in the areas of problem description, model description and assumptions, equations usage and presentation, and English writing style. The submission is hard to read and does not have a clear flow of thought and direction. The submission firstly must be thoroughly read and rewritten by authors to make it more readable. I hope that the authors can use some of the following comments given in this review to improve the submission.

Long sentence of Abstract in lines 14-20 needs to be rewritten to make it more understandable, also break it into more than 1 sentence. What do you mean by “risk loss cost” (line 16)?

Line 36, please change “storage” to “storage facilities”.

Please write an introduction sentence before three bullets starting from line 122.

Authors should briefly describe the meaning of individual costs at the end of Section 1 (daily supervision cost, risk loss cost, risk compensation cost, warehousing cost, warehousing supervision cost).

Please extend the sentence at line 149 “Assume that the costs are based on the unit storage of hazardous chemicals.”. Give us more details on how does hazardous chemicals unit of storage impacts costs.

In line 122 authors wrote “Firstly, the hazardous chemicals warehousing enterprises in the lower level model determine the comprehensive risk level …”. On the other hand, in the Abstract in line 14 authors have different order of priorities “Firstly, in the upper level, the regulatory authorities represented by the government set the penalty coefficient …”. Please be consistent on what is first, second, third in your approach.

The objective function (1) is missing parentheses in order to include all variables in the sum by r. Please make an overview (e.g. list with explanation and definition, or in a form of a table) of all coefficients, parameters, sets, variables, and indexes used in your approach before introducing the objective function and constraints. Make a clear statement what are decision variables in your approach. Also, what type of mathematical formulation do you use (MIP, MILP, MINLP, …), please elaborate on this in the paper.

At line 163 authors wrote “Zr is the daily supervision cost of each unit of goods on behalf of the government and the third-party regulatory agency; Xr is the storage quantity of hazardous chemicals in the warehouse r; Wr is whether the representative stores hazardous chemicals in the warehouse r; Ur represents the average risk loss of the third-party regulatory agency and Gr is the risk compensation cost.” This is an example of a bad English writing style, but also an example of inconsistency. If Zr is the daily supervision cost of each unit of goods then that means that r is an index of a certain good. But later, you wrote that Xr is the storage quantity of hazardous chemicals in the warehouse r, which means that r is an index of a warehouse. Authors need to proof check the entire paper regarding English writing style and mathematical model description.

Equations are not correctly or completely defined, e.g. in (2) Er is outside of the sum by r. What is the range of r in (3) (probably you need to add for al r in R but this must be defined by the authors and should not be left to the interpretations of a reader)? In eq. (4) you have Er equal to the sum of er by r, this is mathematically and logically incorrect: first you lack the range of r in Er in order for this equation to be completely defined; secondly all individual warehouses have the same additional loss expressed as the sum of all warehouses extra cost incurred (E1=sum er, E2 = sum er, etc.). A similar problem is also present in the eq. (9). And these incorrect or incomplete definitions continue throughout the entire model. Please check the entire model and make adequate corrections in all equations.

Authors should write the PSO algorithm that they use (not general but specific with parameters for this case), before Figure 2. Mutation operation is carried on the global best solution or local best solution, not on the global or local optimum (line 240). If you had a global optimum you would not need any further calculations or mutations. On the other hand, how can you tell if the global optimum is obtained by PSO which is a metaheuristic approach by itself? Please make adequate changes to the similar optimum and optimal formulations in the entire paper. Reference [26] cannot be found online (search by doi number and by title gives no such reference)? All references in the literature should be checked.

Figure 2 algorithm is too descriptive, please write it using parameters presented in the paper, if necessary, in more detail. Also, there is again problematic usage of the optimum parse in Figure 2. For example, in the last IF block, you wrote “Whether the maximum number of iterations is reached or the optimal solution satisfying the accuracy requirement is found”. How do you know if the solution is optimal to end the IF loop? Also, if you reached the maximum number of iterations which leads to the end of the calculations, then you did not reach the optimal solution. Please elaborate on this.

It is unusual to use the same notation for different parameters (e.g. authors used R for the total number of warehouses and for the quantitative value of risk level). Please correct this, and check the entire paper for similar errors.

The authors should improve the description of 3.2.1 section. Particularly, can you please improve the description of relations between text in the first paragraph (lines 271-278), equations (21,22) and Tables 1-5 (please use notation in eq. 21 and 22 to connect everything). In equation (22) what does Mr stands for? Also, this equation is not correctly defined (the same issue as in the case of eq. 4).

Figure 4 legend is incomplete.

In conclusion, the authors wrote starting from line 364 “On the basis of summarizing the existing research results, this study analyzes the cost composition of the government, third-party regulatory agencies, the public and the hazardous chemicals warehousing enterprises”. Where exactly did you determine costs of the government, third-party regulatory agencies? Where did you determine the cost of a warehouse facility? What does the “Overall enterprise cost” from Table 7 represents, in other words, what is an enterprise in your case, is it just a hazardous chemicals warehouse facility?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors did improve the submission in some segments. However, the main issues are not resolved. The submission has significant drawbacks in the areas of problem description, model description and assumptions, equations usage and presentation, and English writing style. The submission is still hard to comprehend. Additionally, some important remarks from my previous review round are neglected. Therefore, my recommendation is to reject the submission.  I hope that the authors can use some of the following comments given in this review to improve the submission.

Regarding my previous comment on the coefficients, parameters, sets, variables, and indexes: where did the authors make a clear statement on what is a variable and what is a coefficient or parameter? And if something is a variable, what kind of variable is it? For instance:

·        is Xr a variable that takes value 0 if zero units of warehouse area stores hazardous chemicals (in other words it does not store any hazardous chemicals)?

·        is Wr a binary variable that takes value 0 if warehouse r does not store any hazardous chemicals?

·        if the above is correct, why do you need Wr (by Xr you can clearly define if there are or there are not any hazardous chemicals in a warehouse r)?

English writing style is still at a low level, e.g. “Where: Zr represents the daily supervision cost of goods per unit area in the warehouse r on behalf of the government and the third-party regulatory agency; Xr is the area where the hazardous chemicals were stored in the warehouse r; Wr is whether the representative stores hazardous chemicals in the warehouse r …”. It is unclear from this statement what do you mean by Zr, what is “supervision cost of goods per unit area”? Authors use present and future tenses, e.g. why did you write “where the hazardous chemicals were stored” and after you write “Wr is whether the representative stores hazardous chemicals”? The segment of the sentence “Wr is whether the representative stores” is written badly etc. This paper requires significant improvement in the English writing style.

Regarding the writing style of equations (this was also an issue from my previous review round), it is still not correct. For example, in equation (2), all warehouses have the same Ur1 value because every single Ur1 equals the same sum of parameters from all warehouses in set R (U11= sum_by_r (Er+FrXrWr), U21 = sum_by_r(Er+FrXrWr), etc.). And there are many examples of this in the paper. Authors need to proof check the entire paper regarding academic writing style and mathematical model.

Another remark that was not addressed from my previous review round is the usage of optimal solution terminology in Figure 2. I will repeat the exact comment because it still stands: For example, in the last IF block, you wrote “Reach the maximum number of iterations or satisfying the accuracy requirement”. How do you know if the solution is optimal to end the IF loop? Also, if you reached the maximum number of iterations (or you obtain satisfying accuracy) which leads to the end of the calculations, then you did not reach the optimal solution. Nearly optimal is not optimal, it does not even matter if that error is “microscopical”. If you break the loop after some iterations based on the number of iterations, then you find some solution and you don’t know if it is optimal. The authors' response to this comment is also unusual: “The solution to end the IF loop in this paper refers to the reference [27], and strives to achieve the “approximate” global optimal solution. We have made a quote in the paper, the details will not expand.” Firstly, I don’t understand what will not expand? Secondly, I am not sure if the authors call on reference [27] to justify the usage of the optimal solution terminology in Figure 2 of the submission. If this is the case, I must say that it is not important what reference [27] contains, authors should read what is written in their submission and in Figure 2 and make adequate answers to the raised issue.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

There was a reason for my previous and final recommendation to reject the submission. I will give just a few remarks in the following that confirm my final recommendation to reject the submission.

For instance, the authors' response to the optimality remark is “Although we cannot guarantee that it is absolutely optimal, it is basically guaranteed to reach approximate optimum. Specifically, as in line 258 of the paper, as follow: The termination condition of the algorithm are the maximum number of iterations iterMax=50 ,and the convergence accuracy ...”. Why do you explain in your response that you use the max number of iterations or convergency accuracy to support your claim of optimality? Both of these are proofs that you cannot know if the solution is optimal. You are not researching religious problems. In mathematics 1+1=2 and it does not matter what you believe that is basically guaranteed. And what does basically guaranteed mean regarding optimality? There is no grey zone, either it is optimal in given conditions and constraints or it is not.

If r is set, what is the index of that set? How can a parameter Zr be the cost of warehouse r when r is a set of warehouses?

If Ur and Gr are parameters, why do you minimize both in the objective function (1)? These are “standalone” parts of the objective function (there is no variable associated with them).

If variable Wr=0 and parameter Er=0 (which is a possibility based on the definition of both, both can equal zero based on (3) and (8)) then Ur1=0 and Ur2=0 and Ur=0. On the other hand, the ranges of these are defined in table 1 as Ur>0, Ur1>0, Ur1>0. How is this possible? Also, in (3) you state that Er=0 if dr=0, but in table 1 you state Er>0. How is this possible?

In table 1 you defined qr≥0. You have no upper bound to this risk. If this value takes for example value 2 you will have negative risk Ur (see equation (5)).

…

Again, I still hope that the authors can use some of the following comments given in this review to improve their research and scientific approach in the future.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf