Potential Benefits of Demand Responsive Transport in Rural Areas: A Simulation Study in Lolland, Denmark
Round 1
Reviewer 1 Report
The manuscript presents a current topic for the definition of complementary and useful transport systems for the spread of sustainable mobility and the reduction of the use of private vehicles.
There are some grammatical errors in the text.
It is considered appropriate to underline the novelty of the research in the introductory part, also inserting the problem of modal choices in the current pandemic phase and the definition of the new areas with low demand.
Therefore we recommend reading the following research papers:
1) Papanikolaou, A., & Basbas, S. (2021). Analytical models for comparing Demand Responsive Transport with bus services in low demand interurban areas. Transportation Letters, 13(4), 255-262.
2)Abdullah, M., Ali, N., Shah, S. A. H., Javid, M. A., & Campisi, T. (2021). Service Quality Assessment of App-Based Demand-Responsive Public Transit Services in Lahore, Pakistan. Applied Sciences, 11(4), 1911.
It is necessary to enter acronyms in extended form when they appear for the first time in the text.
It is necessary to enter the sources from which the maps used in the manuscript were acquired.
What type of DRT service is implemented in the examination area? minibus or what else?
Can all the different types of users / social groups use the DRT service?
It is considered appropriate to better justify the choice of the case study (area).
It is necessary to insert a greater explanation in support of figure 6 and figure 7 and 8.
It is considered appropriate to check the coloring of the axes of the aforementioned figures specifying in this sense the use of the different colors.
It is necessary to add the limitations of the research in the final part of the manuscript, underlining whether the current methodology can be applied to other contexts
Author Response
Thank you for the detailed and thorough review.
There are some grammatical errors in the text.
We have reviewed the document and improved the text. If you find that the language needs further improvement, we will request the help of language editing service in the university. That would take approximately 2-3 weeks.
It is considered appropriate to underline the novelty of the research in the introductory part,
The contribution is added in the end of the introduction sections.
also inserting the problem of modal choices in the current pandemic phase and the definition of the new areas with low demand.
There seems to be a potential in DRT in the situation of pandemic. From the side of the operator, the booking component of DRT could be used for controlling occupancy in the vehicles. From the point of view of travellers, DRT could be seen as a safer alternative than a regular bus, again due to the guaranteed occupancy level. However, we have not been able to identify relevant literature to back up such claims.
Also, as you hinted in the question, due to restrictions and teleworking, demand (as we all observed) decreases, and more of “low demand” areas appear. This, in combinations with the previous hypothesis, opens up new (more) possibilities for DRT. This supports to our simulation experiments in section 5.2, where we vary the demand. A reduced demand scenario could be seen as a response to a hypothetical travel or activity restrictions.
We think this is a valid suggestion, but there is not enough of background to speculate in depth on the “new” mode choice and discuss the areas where DRT could have a potential.
We add a short discussion of possibilities of DRT in the pandemic situations into the discussion section.
It is necessary to enter acronyms in extended form when they appear for the first time in the text.
Added explanation of QoS.
It is necessary to enter the sources from which the maps used in the manuscript were acquired.
Added sources as footnotes.
Therefore we recommend reading the following research papers:
1) Papanikolaou, A., & Basbas, S. (2021). Analytical models for comparing Demand Responsive Transport with bus services in low demand interurban areas. Transportation Letters, 13(4), 255-262.
2)Abdullah, M., Ali, N., Shah, S. A. H., Javid, M. A., & Campisi, T. (2021). Service Quality Assessment of App-Based Demand-Responsive Public Transit Services in Lahore, Pakistan. Applied Sciences, 11(4), 1911.
Thank you for the references. We used this as an inspiration to add a bit of discussion about preferences of travellers when it comes to DRT.
What type of DRT service is implemented in the examination area? minibus or what else?
Added description of this into the beginning of “Simulation scenarios and results”. We used 8-seat minibuses.
Can all the different types of users / social groups use the DRT service?
Anyone previously using PT would be using DRT in our setup. We assume that exactly the same people are using both services. We do not have any data on individuals in the input dataset. And we did not model anything but the OD pair and time stamp for travellers.
It is considered appropriate to better justify the choice of the case study (area).
Added more explanation to “Area, data and data processing”. In short: interest from the transport actors, availability of data, convenient location, and representative area.
It is necessary to insert a greater explanation in support of figure 6 and figure 7 and 8.
The section 5 (where the figures are located) is expanded and more explanations included.
It is considered appropriate to check the coloring of the axes of the aforementioned figures specifying in this sense the use of the different colors.
The colours of the axis in Figure 7 correspond to the line colour on the charts. Together with the legends separated into two parts, we were trying to visually highlight what axis corresponds to what line. That is also explained in the caption.
In other figures, colour encodes the belonging of chart to bus (blue) or to DRT (orange) category and it was not possible to colour-code the axis as well.
It is necessary to add the limitations of the research in the final part of the manuscript, underlining whether the current methodology can be applied to other contexts
Discussion section is expanded highlighting more limitations.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The paper “Potential benefits of demand responsive transport in rural areas; simulation study in Lolland, Denmark “ well covers background theory and past studies about DRT system. The authors developed their own open-source tool and applied it to the real data of the rural area in Denmark. Then, they simulated three different scenarios to explore the efficiency and effectiveness of DRT system in the study area. Overall, the paper explains the method development and its process well, and offers a detail explanation on the results of experiments. However, there are some weaknesses in the current manuscript.
First of all, the authors calculated the modal split based on time window only. As known, the price of service is one of the most powerful factors when users select their mode of trips. It would be better if this study could add one more scenario based on the rate difference between DRT and PT.
Secondly, I’m wondering how the authors setup the origin locations of DRT trips. As I understand, they used OD matrix achieved from a transit operator, and then its routes may be similar with normal transit route. If it is correct, it would be hard to reflect the one of the most benefits of DRT service: a new flexible mode covering the first/last-mile.
Finally (and more basically), waiting time of DRT (Figure 5(b)) are unrealistic, and like the authors mentioned, the algorithm they developed should be amended. Otherwise, they may have to show the deviation of the users’ preference of waiting time of current PT in the study area.
In addition, there is editorial error that should be edited from the paper as follows:
- Line 420 in page 12: Figures 4a and 4b -> 4c and 4d
Author Response
Thank you for such an in-depth review and the suggestions.
First of all, the authors calculated the modal split based on time window only. As known, the price of service is one of the most powerful factors when users select their mode of trips. It would be better if this study could add one more scenario based on the rate difference between DRT and PT.
We agree that DRT service would inevitably affect the demand, and cost is indeed one of the important factors for travellers (or an important instrument for policy makers). Prediction of a modal split after introduction of DRT is a great suggestion for future work. However, such prediction was not the goal of this study. With the current simulation approach and time-windows model, it is rather hard to make such prediction. The main part of this study is based on the assumption that the travellers perceive DRT as the same service quality as buses. To achieve that, we are trying to provide the same trip-level characteristic and implicitly assume that ticket cost stays the same.
Introduction of a new parameter (such as trip cost) into the model is quite problematic as it would require modelling the mode choice, which in turn would require modelling the full demand (for which we have not the required data). Additionally, this does not help much in the analysis of operation costs and environmental impact of the DRT service. In this study, we strive to analyse in which situations DRT is relevant. The results of our study could allow to estimate the service costs, if the demand is given as an input.
Secondly, I’m wondering how the authors setup the origin locations of DRT trips. As I understand, they used OD matrix achieved from a transit operator, and then its routes may be similar with normal transit route. If it is correct, it would be hard to reflect the one of the most benefits of DRT service: a new flexible mode covering the first/last-mile.
Yes, what we simulated is a stop-to-stop type of DRT (i.e. DRT drives travellers from bus or train stops to train or bus stops). We would like to note that this type of service is also relevant to consider, as it is known to reduce the operational costs.
It is correctly noted that the paths of DRT vehicles would largely repeat the routes of buses. However, this would be a typical case in the rural areas, where alternatives roads are often unavailable. Flexibility of DRT in this case would be mostly utilised for driving exactly where needed (not the whole bus route) and making detours to pick up travellers from a “neighbour bus route”.
Similar to the previous comment, a door-to-door type of DRT would affect the demand (e.g. attract new users from smaller villages not covered by buses or attract new OD pairs between locations that were not directly connected previously).
Finally (and more basically), waiting time of DRT (Figure 5(b)) are unrealistic, and like the authors mentioned, the algorithm they developed should be amended.
The waiting time that we used for the main experiment (±30 minutes) is large indeed. We argue that buses with a headway of 1 hour could be as inconvenient (force similar waiting times) for the travellers. The counter-argument could be that bus travellers adapted their activities to the existing buses and they do not experience large discomfort every day. The experiment in section 5.3 shows the efficiency of DRT with allowed time window down to the size of 15 minutes (+- 7.5 minutes).
We have added some discussion on the problem of reliability of DRT, and how it could affect the demand.
Otherwise, they may have to show3 the deviation of the users’ preference of waiting time of current PT in the study area.
The deviation of departure time on PT from the desired departure time is shown in Figure 5a.
The figures have been slightly modified to make them easier to understand.
Author Response File: Author Response.docx
Reviewer 3 Report
The manuscript has a very up-to-date topic and is well-organized. I found the manuscript significantly contributing to the existing literature on the transportation domain. Therefore, I recommend it for publication. However, I would like to mention two points to be considered by the authors:
(1) Please avoid presenting the results at the end of the introduction section. Instead, please add a paragraph, explaining the organization and structure of the paper.
(2) Please mention the full form of "QoS" the first time you use it.
Author Response
(1) Please avoid presenting the results at the end of the introduction section. Instead, please add a paragraph, explaining the organization and structure of the paper.
The results in the introduction section were reduced. A paragraph with the organisation was added.
(2) Please mention the full form of "QoS" the first time you use it.
QoS was introduced without explanation. Fixed.
Author Response File: Author Response.docx
Reviewer 4 Report
Overall this is a well-written and interesting paper modeling the potential of demand responsive transport in Denmark. A few questions and comments for the authors:
What are the limitations of the data set and simulation model? These should be discussed, as appropriate.
Based on the findings of the paper, are there any recommendations for policy and/or additional research?
Some additional resources that may be helpful for the authors:
https://escholarship.org/content/qt6533b9wk/qt6533b9wk_noSplash_d73f7b230253f0e5c29cd9f137072c2f.pdf?t=mc2sle
https://www.amazon.com/Paratransit-America-Redefining-Mass-Transportation/dp/027595725X
Author Response
Thank you for the review and the references. It indeed seems that Asian countries have more developed general-purpose on-demand services.
What are the limitations of the data set and simulation model?
The limitations are further elaborated in the discussion section.
Based on the findings of the paper, are there any recommendations for policy and/or additional research?
Added this to Discussion section as well.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
It is necessary to pay attention to the positioning of table 1 because the sentences are interrupted.
It is necessary to standardize the formatting of all tables.
In the new version of the manuscript it is not easy to see the changes made, such as the explanation of the QoS and the new references used as an inspiration to add a bit of discussion about preferences of travellers when it comes to DRT.
Author Response
Thank you for the review.
You write that “In the new version of the manuscript it is not easy to see the changes made”. We wonder if you noticed the pdf version with the “track changes”. We add it here as an attachment to make sure that changes are visible. The attached file includes all the changes made from the original submission.
***
First, we would like to address the two articles you suggested. They both are very relevant articles. We have added Abdullah et al. (2021) as a reference when discussing important factors defining the travellers’ satisfaction. See lines 591-597 (here and further on, we refer to the “clean” submission of the revision 2 file without “track changes”). We did not include the Papanikolaou et al. (2021) reference initially. This article uses a similar approach to the other works we mention in section 2 “Related Work” (lines 132-142). This article elaborates on the optimal headway (or a range of headways) for buses and DRT. In this revision, we utilise it as a bridge between the description of the approach and the design of the DRT service that is used in such analytical studies and the main findings of the articles that we want to highlight.
Regarding novelty: At the end of the introduction section, we highlight the main contribution and the novel results. One of the features that distinguish our work is the type of DRT system studied (DRT to PT transfers). While we are not the first studying this, to our knowledge, only a few studies have considered this. Also, after Diana et al. (2007), few simulation studies have considered the environmental aspect. Combined with the first point, we argue that we provide a novel contribution.
Regarding limitations. The discussion section has been expanded and restructured to include more discussion of limitations. See lines 573-646.
Regarding the language: We have submitted the manuscript to the language editing service. It will, however, take some extra time to be processed and we were not able to provide the results of the language editing as part of this revision.
Author Response File: 
Author Response.pdf
Round 3
Reviewer 1 Report
The manuscript still contains some grammatical errors and typos.
It is advisable to include the sources for the maps used.
It is advisable to replace the semicolon with a colon in the title.
Author Response
Proofreading of the text has been done. See the corrections in the attached "track-changes.pdf".
The sources of maps are added as footnotes 10 and 11 in the "clean" resubmitted version.
The semicolon in the title is replaced by a colon, as advised by you and our proofreader.
Author Response File: Author Response.pdf
