Multicriteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative
Round 1
Reviewer 1 Report
The papers presents a research on a Multi-Criteria Network Optimal Problem with shippers heterogeneous preferences for the China Railway Express (CRexpress). The presented work is original and interesting for publishing, however the reviewer considers that some
Major questions
1 - Figures and Tables are not well referred in the paper. This should be indicated as a minor question however the impact in the results section compromises a clear understanding of the work. Some figures are apparently split in blocks as the ones in lines 369/370, 372/373, 375/376, 427/428, 432/433 just to name some.
Figure 7 is apparently broken in two pages and has a margin. Only one approach should be used for all figures.
2 - The authors refer in the abstract and also in the conclusion that the work can be used for suggestion of subsidies from the government. No discussion regarding this issue is presented on section 3 Results.
3 - The proposed model using 8 criteria is innovative. However it was not clear for the reviewer the capability to transfer the criteria Safety, Flexibility, Traceability from the real world into the model. A model regarding percentages is presented and tested, but remains the question on how a percentage is translated between the real world and model for the three criterion mentioned before. (see lines 350 to 352) A question remains, in the case a shippers considers a relevance of 86% in traceability, how can this be achieved in practice by the CRexpress? In the perspective of the reviewer this keeps the proposed approach far from real application until this gap is solved.
4. The presentation of the CRexpress should be revised. The maximum dimension of the subset of each city type should be clearly indicated (lines 169 t0 175). In section 2.1 no information in given regarding the transport Modes. The process of consolidation towards 3 consolidation centers and a single destination (Hamburg) should be presented clearly in section 2.1 (later this is done in line 306). Namely in Figure 1:
- the three corridors k are named Gateway in the legend
- demand is present on all cities set I, J(n,j), but the figures gives the illusion that demand is only presented on nodes i and n
- origin is mentioned only for nodes i in I
5. The concept of origin city is not clear for the author. If the author well understood, the work there is transport demand on three types of cities denoted by i, n and j. and not only at the origin cities i which cannot be chosen as consolidation centers.
Minor questions
1.On line 110 there is a "(" not closed
2. On section 2.2.2. the variable "m" is not defined.
3.No reference to the hardware used to solve the model (line 301)
4.In Table 5 some utilities are negative for Design B, this is not explained in the paper
5.In Table 4 (line 402) some cities are underlined, with is only explained on line 443
6.There should be an initial Table with the transport demand of all cities, regardless the scenario. Also a table with the distances between all cities pairs will improve the readability of the paper.
Author Response
Response to Reviewer 1 Comments
Major questions
Point 1: Figures and Tables are not well referred in the paper. This should be indicated as a minor question however the impact in the results section compromises a clear understanding of the work. Some figures are apparently split in blocks as the ones in lines 369/370, 372/373, 375/376, 427/428, 432/433 just to name some. Figure 7 is apparently broken in two pages and has a margin. Only one approach should be used for all figures.
Response 1: Thank you for this comment.Mistakes above have been corrected in the new version, please see section 3.
Point 2: The authors refer in the abstract and also in the conclusion that the work can be used for suggestion of subsidies from the government. No discussion regarding this issue is presented on section 3 Results.
Response 2:Thank you for this comment. We have added to the discussion section the following explanation:
“The knowledge about the preferences of different user groups and the performance of different designs can be used further by government to establish how subsidies should be allocated to achieve the desired effects. For example, if the objective of a government subsidy scheme is to achieve a greener transport system, the emission reduction obtained with a design, combined with an assumption about market prices, can be used to evaluate the required subsidy to achieve a unit value of emission reduction in the network. Subsidies could be allocated based on these costs.”
Point 3: The proposed model using 8 criteria is innovative. However it was not clear for the reviewer the capability to transfer the criteria Safety, Flexibility, Traceability from the real world into the model. A model regarding percentages is presented and tested, but remains the question on how a percentage is translated between the real world and model for the three criterion mentioned before. (see lines 350 to 352) A question remains, in the case a shippers considers a relevance of 86% in traceability, how can this be achieved in practice by the CRexpress? In the perspective of the reviewer this keeps the proposed approach far from real application until this gap is solved.
Response 3:Thank you for this comment. We have added to the discussion section the following explanation:
“Different shippers may vary in the satisfaction towards the traceability they received from different supply chain suppliers. Therefore, in our paper we use percentage to represent shippers satisfaction towards the traceability. Global Positioning System (GPS) and Radio Frequency Identification (RFID) can help give the shippers more information about their cargos, yet it comes at a compromise of cost. The level of application of GPS and RFID and tracking of cargos is usually higher in road than railway/CR express, therefore shippers will have higher satisfaction towards road than railway/CR express.”
For the percentage of traceability, 80% was considered for road from origin city to consolidation centers and 60% for railway. The traceability of CRexpress from gateway to the single destination is considered as 60%. Moreover, all the values of 8 criteria are normalized into values in [0,1] for further analysis, therefore the main point of the percentage is to explain the comparison between road and railway.
References of using percentages in modelling are added in the new version, please sea line 1159-1161.
Point 4: The presentation of the CRexpress should be revised.
-The maximum dimension of the subset of each city type should be clearly indicated (lines 169 t0 175).
-In section 2.1 no information in given regarding the transport Modes.
-The process of consolidation towards 3 consolidation centers and a single destination (Hamburg) should be presented clearly in section 2.1 (later this is done in line 306). Namely in Figure 1:
- the three corridors k are named Gateway in the legend
- demand is present on all cities set I, J(n,j), but the figures gives the illusion that demand is only presented on nodes i and n
- origin is mentioned only for nodes i in I
Response 4:Thank you for this comment.
-The maximum dimension of the subset of each city type cannot be defined in section 2.1 since it depends on the result of optimization.
-Description of transport modes are added in the new version, please see line 256-266.
-Process of transportation a single destination is added in the new version, please see line 277.
-Only gateway is kept for k in the new version.
-For city n which is not chosen as consolidation centers will also be treated as origin city, please see line 263-264 in the new version.
-Description of city j has be corrected (see line 275-276) and notes has been added under figure1 for the demand of city j, please refer to figure1 in the new version.
Point 5: The concept of origin city is not clear for the author. If the author well understood, the work there is transport demand on three types of cities denoted by i, n and j. and not only at the origin cities i which cannot be chosen as consolidation centers.
Response 5:Thank you for this comment. The demand of three types of cities was mentioned in line 258-259. More explanation has been added in line 263-270.
Minor questions
Point 1: On line 110 there is a "(" not closed
Response 1: Thank you for this comment. “)” has been added in the new version,.
Point 2: On section 2.2.2. the variable "m" is not defined.
Response 2:Thank you for this comment. Declination has been been added in the new version.
Point 3: No reference to the hardware used to solve the model (line 301)
Response 3:Thank you for this comment. Reference has been added in the new version.
Point 4: In Table 5 some utilities are negative for Design B, this is not explained in the paper.
Response 4:Thank you for this comment. We have added the explanation in the new version. The text now reads:
“Note that since the utility of cost, time and emission are negative, the total utility of some cities might be negative in Table 5”
Point 5: In Table 4 (line 402) some cities are underlined, with is only explained on line 443.
Response 5:Thank you for this comment. Cities with underlines are candidates for consolidation centers, in other words, cities in city set J.
Point 6: There should be an initial Table with the transport demand of all cities, regardless the scenario. Also a table with the distances between all cities pairs will improve the readability of the paper.
Response 6:Thank you for this comment. Transport demand of all cities could be obtained from reference of Zhao (Zhao, L.; Zhao, Y.; Hu, Q.; Li, H.; Stoeter, J. Evaluation of consolidation center cargo capacity and locations for China railway express. Transp. Res. Pt. e-Logist. Transp. Rev. 2018, 117, 58-81.). Distances between all cities pairs have been listed in the appendix and in section 3.1.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper presents a multi-criteria model for mode/route selection in logistics operations. Although the paper has the potential for publication, it is very difficult to review it at its current form as it, unfortunately, seems to be an incomplete effort. The results section has a number of issues that need to be addressed. Some of these are:
- The same bar chart appears multiple times.
- There are incomplete sentences.
- References to tables/figures are missing.
- There are references to tables that do not exist.
Other areas that require improvement are:
- A very thorough proofreading is required as there are multiple grammar mistakes throughout.
- Section 2.3 has several series of equations with small differences between successive series. This makes the section rather long for its purpose. A rationalisation is necessary to minimise the number of equations and only keep the ones that represent the final version of the proposed model.
Author Response
Point 1: The results section has a number of issues that need to be addressed. Some of these are:
- The same bar chart appears multiple times.
- There are incomplete sentences.
- References to tables/figures are missing.
- There are references to tables that do not exist.
Response 1: Thanks for the comments. Mistakes above have been corrected in the new version.
Point 2: Other areas that require improvement are:
- A very thorough proofreading is required as there are multiple grammar mistakes throughout.
- Section 2.3 has several series of equations with small differences between successive series. This makes the section rather long for its purpose. A rationalisation is necessary to minimise the number of equations and only keep the ones that represent the final version of the proposed model.
Response 2:Thank you for this comment.
-We have corrected the grammar mistakes.
-Equations in section 2.3 explains the process of transforming the bi-level model using KKT conditions, therefore the equations have to be kept for references of the one-to-one match between constraints and conditions. We have added the final version of the proposed model , please refer to line 408-420.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper “Multi-criteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative” has been submitted to the journal “Sustainability” [Manuscript ID: sustainability-924830]. This work demonstrates the importance of incorporating shippers’ preference heterogeneity into the optimization of the China Railway express network by a model including price, time, reliability, frequency, safety, flexibility, traceability and emission. There are huge margins of novelty and innovation here. The paper reads also smoothly in terms of language and method, so that it has been interesting to inspect it.
My overall judgement is in the positive. I address the authors only a couple of minor points that could interest the careful readers of this journal. The first one regards the reliability of the data obtained. To the best of my knowledge, this is a major point in any freight survey [please see: Agostino Cappelli, Silvio Nocera. Freight modal split models: data base, calibration problem and urban application. WIT Transactions on the Built Environment Volume 89, 2006, Pages 369-375; María Feo, Raquel Espino, Leandro García. A stated preference analysis of Spanish freight forwarders modal choice on the south-west Europe Motorway of the Sea. Transport Policy Volume 18, Issue 1, 2011, Pages 60-67; José Geraldo Vidal Vieira, Jan C. Fransoo, Carla Deguirmendjian Carvalho. Freight distribution in megacities: Perspectives of shippers, logistics service providers and carriers. Journal of Transport Geography Volume 46, 2015, Pages 46-54]. Yet, the issue of the reliability of the data surveyed and its implications on the model precision have not been discussed. I would briefly comment on both these points at the end of the section 1.
A second point is about the emission specification. CO2 is generally not included in the primary pollutants, but is known as a major driver for climate change. Still it has been chosen as the only pollutant in the model. This choice is polar, and needs to be thoroughly explained [see: L. Int Panis, C. Beckx, S. Broekx, I. De Vlieger, L. Schrooten, B. Degraeuwe, L. Pelkmans. PM, NOx and CO2 emission reductions from speed management policies in Europe. Transport Policy Volume 18, Issue 1, 2011, Pages 32-37; M.S. Alam, H. Perugu, A. McNabola. A comparison of route-choice navigation across air pollution exposure, CO2 emission and traditional travel cost factors. Transportation Research Part D: Transport and Environment Volume 65, 2018, Pages 82-100].
Final point. Please note that in the submitted version all the connections with the references have been lost from section 2 on - a point that should be fixed in a possible resubmission.
Author Response
Point 1: The first one regards the reliability of the data obtained. To the best of my knowledge, this is a major point in any freight survey [please see: Agostino Cappelli, Silvio Nocera. Freight modal split models: data base, calibration problem and urban application. WIT Transactions on the Built Environment Volume 89, 2006, Pages 369-375; María Feo, Raquel Espino, Leandro García. A stated preference analysis of Spanish freight forwarders modal choice on the south-west Europe Motorway of the Sea. Transport Policy Volume 18, Issue 1, 2011, Pages 60-67; José Geraldo Vidal Vieira, Jan C. Fransoo, Carla Deguirmendjian Carvalho. Freight distribution in megacities: Perspectives of shippers, logistics service providers and carriers. Journal of Transport Geography Volume 46, 2015, Pages 46-54]. Yet, the issue of the reliability of the data surveyed and its implications on the model precision have not been discussed. I would briefly comment on both these points at the end of the section 1.
Response 1: Thank you for this comment and the references to other surveys. We are aware of the general limitations of surveys. As the survey itself is not part of our work we refer to the source [57] for details. At the same time we have extended the explanation of the survey to be clearer about the sample and the population it represent. The text now reads:
“In their study, a stated preference survey was conducted from March 29th, 2019 to April 18th, 2019 via face-to-face personal interviews at CHENGDU INTERNATIONAL RAILWAY SERVICE CO., LTD in China. The sample consisted of 63 shippers that use the CRE, which has a customer base of around 3500. This is the largest survey to date of service preferences in this market. Several models were estimated to obtain weights for the criteria.”
Point 2: A second point is about the emission specification. CO2 is generally not included in the primary pollutants, but is known as a major driver for climate change. Still it has been chosen as the only pollutant in the model. This choice is polar, and needs to be thoroughly explained [see: L. Int Panis, C. Beckx, S. Broekx, I. De Vlieger, L. Schrooten, B. Degraeuwe, L. Pelkmans. PM, NOx and CO2 emission reductions from speed management policies in Europe. Transport Policy Volume 18, Issue 1, 2011, Pages 32-37; M.S. Alam, H. Perugu, A. McNabola. A comparison of route-choice navigation across air pollution exposure, CO2 emission and traditional travel cost factors. Transportation Research Part D: Transport and Environment Volume 65, 2018, Pages 82-100].
Response 2:Thank you for this comment. Our assumption is that CO2 is a sufficient proxy for CO2e to compare relative performance levels of road and rail on a [0,1] scale. We agree that this is a limitation concerning the positioning of the modes on this scale. For transparency, therefore, we have added a more detailed explanation about our assumption in the text from line 360. The assumption will not affect our results however as (1) emission valuations are the lowest from all criteria - an order of magnitude lower than the main group of criteria and, more generally, (2) the main conclusions are methodological.
Point 3: Final point. Please note that in the submitted version all the connections with the references have been lost from section 2 on - a point that should be fixed in a possible resubmission.
Response 3:Thank you for this comment.The mistakes in section 2 have been corrected in the new version.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I reviewed the first version of the paper "Multi-criteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative".
Regarding the major point 2 the authors letters mentions an update of the paper that has not been added to the final version (section 4. Discussion and Conclusions).
The minor comments have been addressed and corrected.
It is difficult to follow the updates since the lines pointed for each update are not correct (please check answer to point 3 lines 1156 and to point 4 lines 256 which is not related just to name some).
Author Response
Response to Reviewer 1 Comments
Point 1: Regarding the major point 2 the authors letters mentions an update of the paper that has not been added to the final version (section 4. Discussion and Conclusions).
Response 1: Thank you for this comment.I have added thediscussion in section 4.
Point 2: It is difficult to follow the updates since the lines pointed for each update are not correct (please check answer to point 3 lines 1156 and to point 4 lines 256 which is not related just to name some).
Response 2: Thank you for this comment.Please see the corrected lines below.
Point 3:The proposed model using 8 criteria is innovative. However it was not clear for the reviewer the capability to transfer the criteria Safety, Flexibility, Traceability from the real world into the model. A model regarding percentages is presented and tested, but remains the question on how a percentage is translated between the real world and model for the three criterion mentioned before. (see lines 350 to 352) A question remains, in the case a shippers considers a relevance of 86% in traceability, how can this be achieved in practice by the CRexpress? In the perspective of the reviewer this keeps the proposed approach far from real application until this gap is solved.
Response 3:Thank you for this comment.
“We have added to the discussion section the following explanation:
“Different shippers may vary in the satisfaction towards the traceability they received from different supply chain suppliers. Therefore, in our paper we use percentage to represent shippers satisfaction towards the traceability. Global Positioning System (GPS) and Radio Frequency Identification (RFID) can help give the shippers more information about their cargos, yet it comes at a compromise of cost. The level of application of GPS and RFID and tracking of cargos is usually higher in road than railway/CR express, therefore shippers will have higher satisfaction towards road than railway/CR express.”
For the percentage of traceability, 80% was considered for road from origin city to consolidation centers and 60% for railway. The traceability of CRexpress from gateway to the single destination is considered as 60%. Moreover, all the values of 8 criteria are normalized into values in [0,1] for further analysis, therefore the main point of the percentage is to explain the comparison between road and railway.”
References of using percentages in modelling are added in the new version, please see line 1025-1026.
Point 4: The presentation of the CRexpress should be revised.
-The maximum dimension of the subset of each city type should be clearly indicated (lines 169 t0 175).
-In section 2.1 no information in given regarding the transport Modes.
-The process of consolidation towards 3 consolidation centers and a single destination (Hamburg) should be presented clearly in section 2.1 (later this is done in line 306). Namely in Figure 1:
- the three corridors k are named Gateway in the legend
- demand is present on all cities set I, J(n,j), but the figures gives the illusion that demand is only presented on nodes i and n
- origin is mentioned only for nodes i in I
Response 4:
Thank you for this comment.
-The maximum dimension of the subset of each city type cannot be defined in section 2.1 since it depends on the result of optimization.
-Description of transport modes are added in the new version, please see line 437-438.
-Process of transportation a single destination is added in the new version, please see line 449.
-Only gateway is kept for k in the new version.
-For city n which is not chosen as consolidation centers will also be treated as origin city, please see line 435-436 in the new version.
-Description of city j has be corrected (see line 435-437) and notes has been added under figure1 for the demand of city j, please refer to figure1 in the new version.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The authors have done substantial improvements based on the previous review. However, the following minor changes will benefit the paper:
- A rationalisation of which equations need to appear in their full form must be made. At the moment several equations (ie eq.13 = eq.23, eq.14=eq.25, etc.) appear more than once in the text. There is really no need for that, as the authors can refer to the single equation (instead of copying it again) if necessary.
- Another proofreading round will benefit the paper. Some examples of corrections are given below, but please note that this list is not exhaustive.
- line 347, ....focus of interest.....
- line 484, ....that our proposed model.....
- line 521-523, sentence needs rephrasing as it is not clear what the authors mean.
Author Response
Response to Reviewer 2 Comments
Point 1: The authors have done substantial improvements based on the previous review. However, the following minor changes will benefit the paper:
Response 1: Thank you for this comments.
Point 2:A rationalisation of which equations need to appear in their full form must be made. At the moment several equations (ie eq.13 = eq.23, eq.14=eq.25, etc.) appear more than once in the text. There is really no need for that, as the authors can refer to the single equation (instead of copying it again) if necessary.
Response 2: Thank you for coming back to this points. I really appreciate your advice but so far I think its better to repeat the whole equations because by adding dual variables for each constraint, new equations have to be claimed as the KKT conditions. For more information, we would also suggest the reference “Yang, D.; Pan, K.; Wang, S. On service network improvement for shipping lines under the one belt one road initiative of China. Transp. Res. Pt. e-Logist. Transp. Rev. 2017, S1366554517302909”.
Point 3: Another proofreading round will benefit the paper. Some examples of corrections are given below, but please note that this list is not exhaustive.
line 347, ....focus of interest.....
line 484, ....that our proposed model.....
line 521-523, sentence needs rephrasing as it is not clear what the authors mean.
Response 3:Thank you for this comments and the new version has already gone through a proofreading. We hope it's better now and we also have the certification of English editing in the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
A revised version of the paper “Multi-criteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative” has been submitted to the journal “Sustainability” [Manuscript ID: sustainability-924830R1]. I have reviewed a first version of this paper already, formulating a couple of comments about (i) the reliability of the data obtained from freight surveys, and (ii) the absence of primary pollutants in the model proposed by the authors. I had also suggested the authors some literature to be consulted, in order to clear both of these points.
Both of my comments have been entirely disregarded in their resubmission, so that I think that this paper still needs major revisions
Author Response
Thank you for coming back to these points.
Most respectfully, we disagree with your statement that we have disregarded your comments. We have explained in detail the survey approach and the factors that determine its reliability. Also we have explained the chosen focus of the model regarding emissions, arguing that the suggested extension has little relevance for our conclusions. We repeat these explanations for your convenience below and look forward to receive a more specific reply if you feel that these changes are insufficient.
Point 1: The first one regards the reliability of the data obtained. To the best of my knowledge, this is a major point in any freight survey [please see ...]. Yet, the issue of the reliability of the data surveyed and its implications on the model precision have not been discussed. I would briefly comment on both these points at the end of the section 1.
Thank you for this comment and the references to other surveys. We are aware of the general limitations of surveys. As the survey itself is not part of our work we refer to the source [57] for details. At the same time we have extended the explanation of the survey to be clearer about the sample and the population it represent. The text now reads:
“In their study, a stated preference survey was conducted from March 29th, 2019 to April 18th, 2019 via face-to-face personal interviews at CHENGDU INTERNATIONAL RAILWAY SERVICE CO., LTD in China. The sample consisted of 63 shippers that use the CRE, which has a customer base of around 3500. This is the largest survey to date of service preferences in this market. Several models were estimated to obtain weights for the criteria.”
Point 2: A second point is about the emission specification. CO2 is generally not included in the primary pollutants, but is known as a major driver for climate change. Still it has been chosen as the only pollutant in the model. This choice is polar, and needs to be thoroughly explained].
Thank you for your comment. Our assumption is that CO2 is a sufficient proxy for CO2e to compare relative performance levels of road and rail on a [0,1] scale. We agree that this is a limitation concerning the positioning of the modes on this scale. For transparency, therefore, we have added a more detailed explanation about our assumption in the text from line 360. The assumption will not affect our results however as (1) emission valuations are the lowest from all criteria - an order of magnitude lower than the main group of criteria and, more generally, (2) the main conclusions are methodological.
Round 3
Reviewer 1 Report
I leave the decision to the editor.
Author Response
Please see paper.
Reviewer 3 Report
A third version of the paper “Multi-criteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative” has been submitted to the journal “Sustainability” [Manuscript ID: sustainability-924830R2]. I have reviewed two versions of this paper already. In my most recent comment, I was highlighting for the second time comments about (i) the reliability of the data obtained from freight surveys, and (ii) the absence of primary pollutants in the model proposed by the authors. I had also suggested the authors some literature to be consulted, in order to clear both of these points.
I still feel that both of my comments have not been addressed correctly by the authors. The paper has however some margins of innovation, that make it possibly worth of some attention from this journal. I play Switzerland here and leave the Editor the decision about its publication
Author Response
Comments:
A third version of the paper “Multi-criteria Intermodal Freight Network Optimal Problem with Heterogeneous Preferences under Belt and Road Initiative” has been submitted to the journal “Sustainability” [Manuscript ID: sustainability-924830R2]. I have reviewed two versions of this paper already. In my most recent comment, I was highlighting for the second time comments about (i) the reliability of the data obtained from freight surveys, and (ii) the absence of primary pollutants in the model proposed by the authors. I had also suggested the authors some literature to be consulted, in order to clear both of these points.
I still feel that both of my comments have not been addressed correctly by the authors. The paper has however some margins of innovation, that make it possibly worth of some attention from this journal. I play Switzerland here and leave the Editor the decision about its publication.
Response: Thank you for coming back to these points.
Most respectfully, we disagree with your statement that we have disregarded your comments. We have explained in detail the survey approach and the factors that determine its reliability. Also we have explained the chosen focus of the model regarding emissions, arguing that the suggested extension has little relevance for our conclusions.
Point 1: The first one regards the reliability of the data obtained. To the best of my knowledge, this is a major point in any freight survey [please see: Agostino Cappelli, Silvio Nocera. Freight modal split models: data base, calibration problem and urban application. WIT Transactions on the Built Environment Volume 89, 2006, Pages 369-375; María Feo, Raquel Espino, Leandro García. A stated preference analysis of Spanish freight forwarders modal choice on the south-west Europe Motorway of the Sea. Transport Policy Volume 18, Issue 1, 2011, Pages 60-67; José Geraldo Vidal Vieira, Jan C. Fransoo, Carla Deguirmendjian Carvalho. Freight distribution in megacities: Perspectives of shippers, logistics service providers and carriers. Journal of Transport Geography Volume 46, 2015, Pages 46-54]. Yet, the issue of the reliability of the data surveyed and its implications on the model precision have not been discussed. I would briefly comment on both these points at the end of the section 1.
Response 1: Thank you for this comment and the references to other surveys. We are aware of the general limitations of surveys.
As the survey itself is not part of our work we refer to the source [55] for details, which has been accepted last week for Transportation Research PART A.
At the same time we have extended the explanation of the survey to be clearer about the sample and the population it represent. The text now reads:
“In their study, a stated preference survey was conducted from March 29th, 2019 to April 18th, 2019 via face-to-face personal interviews at CHENGDU INTERNATIONAL RAILWAY SERVICE CO., LTD in China. The sample consisted of 63 shippers that use the CRE, which has a customer base of around 3500. This is the largest survey to date of service preferences in this market. Several models were estimated to obtain weights for the criteria.”
Point 2: A second point is about the emission specification. CO2 is generally not included in the primary pollutants, but is known as a major driver for climate change. Still it has been chosen as the only pollutant in the model. This choice is polar, and needs to be thoroughly explained [see: L. Int Panis, C. Beckx, S. Broekx, I. De Vlieger, L. Schrooten, B. Degraeuwe, L. Pelkmans. PM, NOx and CO2 emission reductions from speed management policies in Europe. Transport Policy Volume 18, Issue 1, 2011, Pages 32-37; M.S. Alam, H. Perugu, A. McNabola. A comparison of route-choice navigation across air pollution exposure, CO2 emission and traditional travel cost factors. Transportation Research Part D: Transport and Environment Volume 65, 2018, Pages 82-100].
Response 2:Thank you for this comment. Our assumption is that CO2 is a sufficient proxy for CO2e to compare relative performance levels of road and rail on a [0,1] scale. We agree that this is a limitation concerning the positioning of the modes on this scale. For transparency, therefore, we have added a more detailed explanation about our assumption in the text from line 360. The assumption will not affect our results however as (1) emission valuations are the lowest from all criteria - an order of magnitude lower than the main group of criteria and, more generally, (2) the main conclusions are methodological.
