An Economic Optimization Model of an E-Waste Supply Chain Network: Machine Learned Kinetic Modelling for Sustainable Production

Comment 1

This paper presented the study of electronic waste recycling based on a multidimensional analysis on cost functions of a framework structured on three modules - environmental, economic, and social uncertainties. Each module is independently classified using machine learning (ML) protocols. From a long list of possible contributors, the model identifies and ranks two key contributors to sustainability: global energy consumption and the volume of carbon dioxide generated. During the investigation, a hybrid AHP-PCA method was developed. Model results were verified using a case study to outline a future E waste-supported roadmap.

I appreciate the way the problem is addressed, but I have a few comments to correct and improve the paper, so I propose the following changes and additions:

Response

Thank you for the comment.

Comment 2

The template must be respected

Response

Thank you for the comment. In the resubmitted version, we have strictly adhered to the journal template.

Comment 3

Reference section: Authors must carefully revise the format of references according to journal instructions

Response

Thank you for the comment. We have rectified the references as required by the journal template.

Comment 4

Make references in the text according to journal instructions;

Response

Thank you for pointing this out. We have converted all the in- text citations into numbered third brackets.

Comment 5

Figures: 5; 7; 8; 9; 10; 11; 12 - complete with a) and b) because there are two pictures;

Response

Thank you for the comment. The figures are now denoted with (a) and (b) and duly mentioned in the figure caption.

Comment 6

The conclusions should be improved, more clearly structures the results on the three pillars of sustainability.

Response

Thank you for the comment. We have completely reworked the conclusion section. The changes are marked in orange-colored lines 694 – 721.

Comment 1

This manuscript may be interesting, but unfortunately there is no hard data, only assumptions and theoretical aspects. This manuscript should be restructured and show a greater amount of data, to be considered for publication.

Response

There is some misunderstanding here. This is a mathematical modeling study that generates data starting from a theoretical construct. This is not a data modeling work in case the reviewer has not noticed this.

Comment 2

in the introduction section there are several paragraphs that use the same reference. It is suggested to use different references.

Response

Thank you for the comment. We have used different references wherever suitable.

Comment 3

the introduction section (section 1) is too long. Authors should restructure the introduction section.

Response

Thank you for the comment. We have revised the entire introduction, targeting precision while still being substantive. We have uncluttered a few items and added research questions and objectives to give it a structure. The changes are marked in orange.

Comment 4

The first paragraph of the materials and methods section is precisely the objective of this work, which is not reflected in the introduction section. It is suggested to restructure those paragraphs.

Response

Thank you for the comment. We have reworked the materials and methods section. The changes are marked in orange.

Comment 5

The study area is not described. The authors indicate that they use anonymous data. With this information, the context of the application of this work is not clear.

Response

Thank you for the comment. Please refer to Line number 211. In the manuscript, we have clearly stated that the data is not just any data, it has been collected from one of the leading Indian e-waste management companies. Due to IP protection, we cannot reveal the name of the company. For follow-up studies and authenticity, the data used in the simulation have been furnished in the supplementary section.

Comment 6

This work poses too many equations and assumptions and few real (hard) data. Some graphs are shown that it is not clear where they come from (fig 5 and 6)

Response

This is a work based on mathematical modeling outcomes of which are validated against data. The current investigation is a follow-up work (DOI: 10.1016/j.physa.2022.128085) of a generic supply chain architecture analyzing profitability against resources. The present study draws from that infrastructure and restructures the model to analyze the e-waste scenario. Given the nature of the work, which is mathematically intensive, it is only expected that it will consist of assumptions and equations.

Please refer to line number 471 –

‘Equations (10) and (13) are solved using data obtained from an anonymous multi-award winner Indian E- waste recycler company. MATLAB R2019a (bvp4c) was used to solve the system of equations concerning the solutions of the corresponding boundary value problems (Table 1).’

Figures 5, 6, etc. represent solutions of a boundary value problem described in equations (10) and (13) that feature the time dynamic evolution of the uncertainty variables. The results strongly connect with qualitative changes observed from real data which can also be quantitatively appraised given appropriate boundary and initial conditions.

Comment 7

Statistically, this work has no strengths since there is no data. It is suggested to show some data to make the manuscript more understandable

Response

Thank you for the comment. This work is not a study in statistics. This is a more detailed and intensive study utilizing tools from mathematics and Machine Learning. Our data come from numerical solutions of the model represented in eqns (10) and (13) and go way beyond spreadsheet plotting of process data. We emphasize that the model validation is based on real life hard data that were painstakingly collected from a leading e-waste management company in India. All of these have been explained in detail in the revised version.