Using Hierarchical Facility Location, Single Facility Approach, and GIS in Carsharing Services

Round 1

Reviewer 1 Report

1) The literature review is long, showing the authors have done a proper job on this.  Some kind of summary, pointing on the most relevant findings would be nice.

2) The title indicates Car Sharing as the main topic. However, the solutions presented are just to a very limited level discussed related to Car Sharing, but rather to general business case location-allocation principles. No discussion on the difference between Car Sharing (discussed in the introduction) and Car Rental (not mentioned in the article). I would have expected some discussion on requirements to e.g. the level of facility service (e.g. capacity, variety of cars) to become attractive, more than the single requirement of 30min walking distance. Could also be relevant to discuss contradicting requirements (e.g. the demand within 30min walking distance is not enough for the needed facility service level)

3) The chapter 2.3 describes the ArcGIS implementations. References to other software implementations than ArcGIS will improve the article.

4) Two GIS principles used for allocation. The first uses distances along a transport network, shown in figure 3. The second uses Voronoi diagrams, i.e. Eucledian distances, shown in figure 4. Cannot see how these two are working together. Should be better explained.

5) In ch 4.2 the car makers service seems to be one set of "centers" of importance, however not candidates for location of car sharing stations (hotels, shopping malls, parking lots and colleges). The table 2 and figure 5 can also be interpreted like this. An explanation on this would help

6) The figure 4 is not possible to read. The amount of details must be reduced and the use of colors must be improved (close to impossible to see the yellow)

7) (important)The chapter 4, especially the figure 5 presents the final result. However, it is not easy for people not familiar with São Paulo to see "the quality" of the result. Some kind of metrics, possibly comparing with other solutions, and/or showing that the initial requirements are fulfilled would be useful. If the result is meant to be the optimal solution, even more metrics are needed.

 

Other comments:

Line 148: Unclear what the word "we" refer to

Line 251: Unclear what is meant by "this article"

Line 306-307: No article from Murray, Church and Feng (2020) in the References. The one from Murray does not mention the Weiszfelt algoritm.

Line 483 - 485: Further editing needed, possibly removing some text.

Author Response

 

• The literature review is long, showing the authors have done a proper job on this. Some kindof summary, pointing on the most relevant findings would be nice.

 

Some additional information has been added to the text, in order to present some conclusions and some of the methodology. Lines – 81-85.

 

In summary, a small conclusion from the results of the two types of methodology applied in the GIS achieved the objective of the article, because an optimal location was chosen for the car markers potential partnership candidates. The two application models chose different establishments in some cases, but the proximity of the establishments are not great distances, belonging to the same district.

 

 

• The title indicates Car Sharing as the main topic. However, the solutions presented are just toa very limited level discussed related to Car Sharing, but rather to general business caselocation-allocation principles. No discussion on the difference between Car Sharing (discussed inthe introduction) and Car Rental (not mentioned in the article). I would have expected somediscussion on requirements to e.g. the level of facility service (e.g. capacity, variety of cars) tobecome attractive, more than the single requirement of 30min walking distance. Could also berelevant to discuss contradicting requirements (e.g. the demand within 30min walking distance isnot enough for the needed facility service level)

 

We agree with the Reviewer and the title was rewritten emphasizing the methodology of location-allocation being applied in GIS, with the objective of solving a station location problem for carsharing.

Using Hierarchical Facility Location, Single Facility Approach and GIS in Carsharing Services.

 

 

• The chapter 2.3 describes the ArcGIS implementations. References to other software implementations than ArcGIS will improve the article.

 

Another software was added, QGis, where it is also possible to perform this same analysis, but with a smaller number of algorithms available compared to ArcGis. Lines – 193-201.

 

The location-allocation methodology is used for different ways to find the best location to solve a given problem. GIS software is used to solve this type of problem, because it has the ability to encompass different types of data and their spatial loca-tion, and the integration of these two features solves the problem in a practical and lo-calized way.

Several GIS software has this location-allocation feature, for example the QGis (http://www.qgis.org) software, which is a free operating system. It features Location Lab QGIS Plugin where it also does some analysis for this purpose. In this article ArcGis was used, because it provides many different kinds of location-allocation algo-rithms and also has a friendly interface.

 

• Two GIS principles used for allocation. The first uses distances along a transport network,shown in figure 3. The second uses Voronoi diagrams, i.e. Eucledian distances, shown in figure4. Cannot see how these two are working together. Should be better explained.

 

It was explained why these two models were used to answer the article's objective. Lines 378-382.

 

Two GIS principles were used for assignment. The first uses distances along a transport network, shown in figure 3. The second uses Voronoi diagrams, i.e. Eucledian distances, shown in figure 4. To compare which model would be more appropriate, or also as was the case in this paper to present whether there is any difference in locality, the result of these two distinct applications, but with a similar result, is presented in Figure 5.

 

 

• In ch 4.2 the car makers service seems to be one set of "centers" of importance, however notcandidates for location of car sharing stations (hotels, shopping malls, parking lots and colleges).The table 2 and figure 5 can also be interpreted like this. An explanation on this would help

 

It was explained why the centers of importance are the car makers. Lines 344-347

 

The car markers were chosen because the company will apply carsharing in a to-tal way, i.e. providing the cars, the maintenance of the cars, and also the total applica-tion of the business. It is important to emphasize that the car markers have no physical space to store these cars for the carsharing application [9].

 

• The figure 4 is not possible to read. The amount of details must be reduced and the use ofcolors must be improved (close to impossible to see the yellow)

 

Figure 4 has been changed. For the sake of clarity, it has been separated into 5 parts. The color yellow has also been changed to purple. Lines 375-376.

 

• (important)The chapter 4, especially the figure 5 presents the final result. However, it is not easy for people not familiar with São Paulo to see "the quality" of the result. Some kind of metrics, possibly comparing with other solutions, and/or showing that the initial requirements are fulfilled would be useful. If the result is meant to be the optimal solution, even more metrics are needed.

 

Figure 5 has been changed. A figure was added with the municipality of São Paulo and where the result is located within the municipality. Lines 389-390.

Table 1 modified. Added the total number of commercial establishments in the municipality and also added the road network. Using the data applied in the article between the years 2015 and 2018. Lines 360-361.

 

 

Other comments:

Line 148: Unclear what the word "we" refer to

Was edited in line 138 - This article presents

Line 251: Unclear what is meant by "this article"

Was edited - was excluded

Line 306-307: No article from Murray, Church and Feng (2020) in the References. The one from Murray does not mention the Weiszfelt algoritm.

Article added in reference

Line 483 - 485: Further editing needed, possibly removing some text.

Was edited - following the journal's referencing model

Author Response File: Author Response.pdf

Reviewer 2 Report

The main purpose of the article is to identify the most advantageous locations of car sharing stations. I find the research problem taken up by the Authors as very important and current. Car sharing is a case of shared mobility service that is becoming more and more popular in urban areas due to its positive impact on the environment by minimizing congestion and reducing the need for parking spaces. The correct location of the stations may affect the efficiency of this system and its adaptation to the expectations of users.

The Authors used seven different types of location-allocation and Voronoi diagrams in combination with optimization algorithms to find the best solution. They adopted the maximization of demand as a criterion for selecting the best location of the car sharing station while minimizing the distance to it. In my opinion, this approach is justified.

The manuscript is written in a consistent, synthetic, and logical manner. The substantive methodological part was well described, and the results were properly presented and commented on. The strong points of the article are also the solid literature review and the use of public free databases in the analyses.

However, I have two minor editorial comments:

- in lines 222-228 the Authors refer to the chart which was not included in the article,

- in my opinion Figure 2 showing the methodological flowchart should be included in the text before discussing the individual phases (e.g. in line 284). Moving the figure earlier would certainly make the article clearer.

Author Response

Reviewer 2:

 In lines 222-228 the Authors refer to the chart which was not included in the article

Was edited - was excluded

 

In my opinion Figure 2 showing the methodological flowchart should be included in the text before discussing the individual phases (e.g. in line 284). Moving the figure earlier would certainly make the article clearer.

Figure 2 has been moved

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The paper is improved in several parts according to first set of comments. That is good.

The figure 4 is greatly improved. Unfortunately a new confusion is introduced by selecting different extent for the five sub-parts of the figure. Same extent would make it easier to see the connection between the five.

In figure 5 a small overview map of the whole city is added, unfortunately without any improvement of the understanding of the content of the figure.

However: I still have problems understanding the conclusions and explanations connected to the presented result in the figure 5 and discussed in the chapter 5. I miss the arguments for the result to be an optimal/good solution. A more clear explanations to the bullet points below could maybe help:

• it is stated (lines 397-398) that the locations chosen are in two named districts. According to the figure 5 many of the dots (the red and yellow) in the map are outside the specified regions. Please explain.
• Maybe my main challenge is not properly understanding the explanation on requirement given in the lines 301 - 306, especially on the walking distance requirement. I expected the important walking distance was for a potential customer ("car renter") to reach the collection/deposit sites. This is not what is written in these lines. In the text it seems like somebody are supposed to walk from a carmaker service to the collection/deposit sites. Please explain.
• The presentation in figure 5 does not show the "system boundaries"/ "covered service area" for the analysis. 
• It is stated in line 382 that the two applications give the same result. But the red and yellow circles in figure 5 shows a different picture. Please explain.
• If I (or somebody else) were to set up a carsharing service in this part of the city, please show clearly (preferably in a map) where I according to the presented analysis, should locate the service sites, including the area of the city covered by the service, number of sites needed and the location of the sites.

 

Minor comment: 

- the text in lines 206 - 233 seems to be copied from some ERSI source. But the line 225 refers to a chart obviously part of the ESRI text, but not copied into the paper. Either delete the reference to the chart, or include the chart in the paper.

Author Response

Answers to Reviewer 1 – 2^nd round

 

We would like to thank the Reviewer for his suggestions, and we hope that this time we have satisfactorily modified the document.

 

We tried to reduce the confusion indicated by bringing back the previous figure and keeping the current one, in the text called figures 4a and 4b. We introduce text at the beginning of item 4.3, in lines 344 to 349. This part of the text, the names of the figures and their references are in red in the new version of the text.

 

A text trying to better explain fig. 5 was inserted in item 3., lines and 285 – 289, and other text in the lines 365-367 before the figure. Both parts are red in the current text.

 

The rental stores and share points considered are those existing in the two selected regions. The objective is experiment is to tell the owner of the carsharing system, the best options for the installation of car delivery and pick-up points, from the perspective of the vehicle supplier. Aspects referring to the users of the system can be seen in [45]. The text was in red in the article.

We agreed with the Reviewer and that part was replaced. Lines 200 – 202 of the new version of the article.

Other minor modifications and the insertion of a new bibliography are also marked in red.

Round 3

Reviewer 1 Report

I still cannot fully understand the initial challenges connected to carsharing adressed in the paper. Calculation of walking distances from car maker service spots (figure 3) still seems irrelevant to me . The cars in a shared pool are picked up by users from the carsharing spots. Visiting car maker service should only be relevant for a "broken car". And in the cases needed, the driving time using a (broken, but still operational) car is relevant, alternativly the car should be picked up by a service truck and transported to the car maker service location.

However, the walking distance from potential car users positions to the pickup spots are relevant, but not discussed at all in the paper.

When not understanding the case to be solved, the presentation of answers become a challenge.

The "GIS handling" of this to me not understandable case, seems still to be ok. And that is possibly the most important part of this paper.