Optimization in Decision Making in Infrastructure Asset Management: A Review
Round 1
Reviewer 1 Report
The paper provides a potentially useful meta-study of decision-making methods in infrastructure asset management. These sorts of studies are useful in fields from time to time. To my knowledge there have not been a lot of them in infrastructure management (one is noted below). However, in my opinion, there are substantial improvements that are required in order to achieve that potential.
In order to locate decision-making within asset management as an intellectual discipline it would be very useful to refer to Daan F.J. Schraven, Andreas Hartmann & Geert P.M.R. Dewulf (2015) Research orientations towards the ‘management’ of infrastructure assets: an intellectual structure approach, Structure and Infrastructure Engineering, 11:2, 73-96, DOI: 10.1080/15732479.2013.848909. This paper from a meta-study provides a means for locating decision-making within the field as a whole.
Definition of asset management (p.1). The definition provided relies almost exclusively on a maintenance and rehabilitation (M&R) basis. While I recognise that this is not at all unusual in engineering-based asset management (as here) there are definitions available within infrastructure engineering approaches that take a broader view and which also allow greater precision in definition. For example, Austroads Guide to Asset Management (2006) defines asset management as being the whole of three parts: capacity change through investment, operations of the network, and M&R. I think that this is a most useful understanding of asset management that is not always taken up in theory or practice. Somewhat unfortunately the most recent (2018) edition does not seem to consistently take this holistic view. What this suggests is that M&R is not the whole story. I don't have a problem if the study is confined to only M&R which may be the case here, but please do not represent M&R as being the whole of infrastructure asset management.
This broader definition of asset management would, in turn, allow greater precision in discussing 'projects and network' aspects. I say that because, to my observation, in the broader asset management there are projects that relate to capacity change - new or upgraded roads or pipelines - and projects that relate to M&R. Network aspects then become a non-project operational domain, It is unclear here whether M&R or capacity change projects are meant here, or both. I suggest that decision-making requirements in each aspect may be different and should be examined as such. This also has implications for the content and discussion of Figure 2.
I also make the observation because I find that much construction-based literature poorly defines 'projects', especially when they talk about 'project life cycle'. Typically a project is a time delimited activity. It has a start, a duration and an end. This makes the inclusion the post-project life cycle problematic though useful in providing an awareness of the longer term effects of the project which otherwise can be downplayed or ignored in a straight project focus. Life cycle really means the operational or use phase of a physical assets life which is separate from the 'project' part..
Enough of the broad, philosophical issues.
p.3 Abelson and Flowerdew is referred to an early piece of literature - date in the text is important in establishing that earliness without having to refer to the reference list. This is even more important when shortly thereafter 'since then' is used - since when?
Figure 1 has several issues. Firstly, it does not sufficiently distinguish between categories when viewed in black and white printing. Acronyms SOO and MOO are not defined, though they are later in the paper. A legend is required here. The statement in the text that the increase in the literature continues from the 2000s to the after 2010 is not supported by the figure itself. The height of the 2000s bars are about the same as the after-2010 bars. The qualification that is required that might justify such a claim is to note that the post-2010s period is not complete at the time of analysis, for instance 7 or 8 years of the decade rather than the full ten years.
Figure 2 has similar issues. See previous discussion about definition of project in asset management - capacity change or M&R. There appears to be a distinct change is slope of the increase up to 2000s and post-2000s that should be remarked on. When discussing the figure's content greater precision is required in, for instance, 'most' and 'a considerable amount' of the literature. The figures sitting behind Figure 2 suggest that these terms should be more precisely defined with precise numbers and/or percentages.
p.5 line 169 'under an annual budget' seems to need 'limitation' or 'constraint' explicitly included rather then being left implied, as here.
p.5 line 175 what 'management benefit' is is not clear. Is it a benefit of or from management or a benefit to management?
Figure 4 requires work. In principle I find pie charts unscientific. Useful in communicating to a non-academic audience but as part of the enduring body of knowledge I find that that are imprecise things. This is especially so when, as here, no figures, say percentages, are included in either the figure or its attendant text. See also comments about including numbers when discussing things like 'most'.
p.5 line 189 - evidence that demand for Pareto solutions is growing? A citation would support this claim.
pp.5 & 6 discussion of more complex Pareto solutions and the reversion to a single solutions suggests, to me, that they, in effect, resort to a SOO form of decision-making. If this is so then this should be commented on.
Citations for heuristic definition. While Silver et al. may have had a definition, Tversky and Kahneman's work on heuristics is much more seminal. It won Kahneman a Nobel Prize in Economics (Tversky's early death prevented his award).
Classification of the 3 types of Applied Optimization Methods. I am not convinced that the software classification is a sufficiently distinct category to warrant its treatment as such. To my mind, the software algorithms uses variants of deterministic and/or heuristic methods. This suggests that much better categorisation and classification justification is required. I'd recommend Philip M. Podsakoff, Scott B. MacKenzie, and Nathan P. Podsakoff (2016) 'Recommendations for Creating Better Concept Definitions in the Organizational, Behavioral, and Social Sciences', Organizational Research Methods. Vol. 19(2), 159-203 as a useful reference in achieving greater conceptual definition. The issue is also relevant to p.6 lines 235-238.
Figure 5 has similar issues to those denoted above for earlier figures. It is unclear what the inner and outer rings denote. This is compounded by black and white reproduction issues but there does not appear to be supporting textual description either. Given previous comments about pie charts I suggest that a table would be more precise. The sequence of the discussion in the text does not follow that implied in the order of the legend items. For instance, Enumeration is first in the legend but appears halfway through the section.
p.6 line 227 expert adjustment to methods does raise all sorts of conceptual issues because while there is an underlying assumption that experts make technically rational decisions. It is just as likely that they use things like expert heuristics (rules of thumb learnt over the course of acquiring their expertise) and possible even (shock, horror!) subjective evaluations.
Check the line of argument to ensure that consistent claims are made. pp.7 & 8 have at least 3 that I can identify. These should be clarified.
line 272 contradicts earlier statements in line 257.
line 288 appears to contradict line 287.
lines 298 & 299 contradicts line 292
p. 8 line 313 Please clarify what is meant by 'financial expense of management'. Is it the expense of the management itself, or is it, more likely from my experience, the financial expense of M&R? lines 316-317 suggest something else again, perhaps, a broader categorisation of 'expenses' though here imprecision is evident. Are 'users' the users of decision-making methods or infrastructure users? Is 'treatment cost' referring to M&R treatments or something else? Does 'agency cost' equate to costs of the (government, usually) agency responsible for the asset management or does it, invoking agency theory refer to a different form of agency?
The last couple of outcome paragraphs - on level of service and other outcomes - appear to be afterthoughts, particularly the last one. Their significance is non commented on, merely listed.
Conclusion, last paragraph. Given the objective of the paper and the commentary about some of the challenges in application, particularly around complexity and scale of decision-making it seems useful to include some consideration of these things in the concluding statements.
Appendix A notes that 337 pieces of literature are examined. However, the list in Appendix A contains only 115 citations. This suggests one of two things. 1. Adjust the 337 to 115 which may also require adjustment to the method description to be consistent. 2. Include all 337 articles. For this kind of meta-study which surveys extensive literature it is my experience that all articles are cited, even though this produces very large reference lists. I note the existence of a second reference list which as far as I can see replicates that in Appendix A. This suggests to me that the problem could be partly solved by have a single list of all relevant references.
Author Response
Dear Reviewer,
Thank you very much for your comments! They are very helpful and valuable. Now the manuscript has been reviewed according to the comments. The responds to all the comments are added as attachment.
Thank you again for your support!
Regards,
Lin
Point 1: The paper provides a potentially useful meta-study of decision-making methods in infrastructure asset management. These sorts of studies are useful in fields from time to time. To my knowledge there have not been a lot of them in infrastructure management (one is noted below). However, in my opinion, there are substantial improvements that are required in order to achieve that potential.
Response 1: Thank you for all the comments! The manuscript is improved accordingly.
Point 2: In order to locate decision-making within asset management as an intellectual discipline it would be very useful to refer to Daan F.J. Schraven, Andreas Hartmann & Geert P.M.R. Dewulf (2015) Research orientations towards the ‘management’ of infrastructure assets: an intellectual structure approach, Structure and Infrastructure Engineering, 11:2, 73-96, DOI: 10.1080/15732479. 2013.848909. This paper from a meta-study provides a means for locating decision-making within the field as a whole.
Response 2: Thank you for the suggestion! This paper discusses the IAM and its topics, while this manuscript mainly focuses on the applications of optimization when making management decisions. Some related finding of this suggested paper is cited in the manuscript (see Section 1).
Point 3: Definition of asset management (p.1). The definition provided relies almost exclusively on a maintenance and rehabilitation (M&R) basis. While I recognise that this is not at all unusual in engineering-based asset management (as here) there are definitions available within infrastructure engineering approaches that take a broader view and which also allow greater precision in definition. For example, Austroads Guide to Asset Management (2006) defines asset management as being the whole of three parts: capacity change through investment, operations of the network, and M&R. I think that this is a most useful understanding of asset management that is not always taken up in theory or practice. Somewhat unfortunately the most recent (2018) edition does not seem to consistently take this holistic view. What this suggests is that M&R is not the whole story. I don't have a problem if the study is confined to only M&R which may be the case here, but please do not represent M&R as being the whole of infrastructure asset management.
Response 3: Thank you for the comment! The authors totally agree that maintenance and rehabilitation is not the whole story of Infrastructure Asset Management. This was clarified in Section 1.
Different organizations have different definition of Infrastructure Asset Management. Examples are added in Section 1 including the definition by Austroads.
Point 4: This broader definition of asset management would, in turn, allow greater precision in discussing 'projects and network' aspects. I say that because, to my observation, in the broader asset management there are projects that relate to capacity change - new or upgraded roads or pipelines - and projects that relate to M&R. Network aspects then become a non-project operational domain, It is unclear here whether M&R or capacity change projects are meant here, or both. I suggest that decision-making requirements in each aspect may be different and should be examined as such. This also has implications for the content and discussion of Figure 2.
Response 4: Thank you for the comment! The authors agree that Infrastructure Asset Management not only focuses on maintenance and rehabilitation (M&R), but a broader range.
(1) This manuscript is based on the reviewed literatures. Most reviewed literatures focus on M&R. Hence M&R becomes a main topic in the manuscript; but it is not the only one. Some literatures also discuss the update work. Clarification is added in Section 3.2.
(2) In this manuscript the project- and network- level is classified according to whether this IAM makes a management strategy for the assets as a whole or makes management strategies for the segments/parts of the assets. This clarified in Section 3.2.
Point 5: I also make the observation because I find that much construction-based literature poorly defines 'projects', especially when they talk about 'project life cycle'. Typically a project is a time delimited activity. It has a start, a duration and an end. This makes the inclusion the post-project life cycle problematic though useful in providing an awareness of the longer term effects of the project which otherwise can be downplayed or ignored in a straight project focus. Life cycle really means the operational or use phase of a physical assets life which is separate from the 'project' part.
Response 5: Thank you for the comment! Yes, researchers may define “project” in different manners. Here the project-level decision making is defined in Section 3.2.
The authors agree with your idea about post-project and life-cycle, while they are beyond the scope of this manuscript.
Point 6: p.3 Abelson and Flowerdew is referred to an early piece of literature - date in the text is important in establishing that earliness without having to refer to the reference list. This is even more important when shortly thereafter 'since then' is used - since when?
Response 6: Thank you for pointing it out! The year is added.
Point 7: Figure 1 has several issues. Firstly, it does not sufficiently distinguish between categories when viewed in black and white printing. Acronyms SOO and MOO are not defined, though they are later in the paper. A legend is required here. The statement in the text that the increase in the literature continues from the 2000s to the after 2010 is not supported by the figure itself. The height of the 2000s bars are about the same as the after-2010 bars. The qualification that is required that might justify such a claim is to note that the post-2010s period is not complete at the time of analysis, for instance 7 or 8 years of the decade rather than the full ten years.
Response 7: Thank you for the comment!
(1) The authors simply thought the figures will be printed in color, so we use different color for different categories. Now it is changed.
(2) SOO and MOO are not used until Section 3.3. Yes it is not wise to leave them in Figure 1. Hence this figure is split into two, one with the total application is in Figure 1 (see Section 3.1) and the other with SOO and MOO is in Figure 4 (see Section 3.3). Please note Figure 4 is a newly added figure.
(3) Figure 1 presents the total number of applications with decades. Because the decade of 2010-2020 is not completed yet at the time of analysis, hence the last bar is not a full bar. Even use 7 or 8 year of decade as coordinate unit, the issue still exists because when the manuscript is drafted, not all the literatures in 2017 and 2018 are published, and the bar is still not full. However, even the last bar is not full, the increase of the application after 2000 is clearly shown in Figure 1.
Point 8: Figure 2 has similar issues. See previous discussion about definition of project in asset management - capacity change or M&R. There appears to be a distinct change is slope of the increase up to 2000s and post-2000s that should be remarked on. When discussing the figure's content greater precision is required in, for instance, 'most' and 'a considerable amount' of the literature.
Response 8: Thank you for the comment! Similar with Figure 1, Figure 2 is also improved (see Section 3.2).
The definition of project is reviewed in Section 3.2.
Also, the description of figures’ content is all improved with more precise words (see the changed in Section 3).
Point 9: The figures sitting behind Figure 2 suggest that these terms should be more precisely defined with precise numbers and/or percentages.
Response 9: Thank you for the comment! Now the percentages are added on the figures.
Point 10: p.5 line 169 'under an annual budget' seems to need 'limitation' or 'constraint' explicitly included rather then being left implied, as here.
Response 10: Thank you! It is changed to “subject to an annual budget”.
Point 11: p.5 line 175 what 'management benefit' is not clear. Is it a benefit of or from management or a benefit to management?
Response 11: Thank you! Management benefit has different definition in different cases. Here the authors only want to give an example of concordant objectives. Now to avoid the confusion, it is changed to “improving the network condition and increasing the level of service” in Section 3.3.
Point 12: Figure 4 requires work. In principle I find pie charts unscientific. Useful in communicating to a non-academic audience but as part of the enduring body of knowledge I find that that are imprecise things. This is especially so when, as here, no figures, say percentages, are included in either the figure or its attendant text. See also comments about including numbers when discussing things like 'most'.
Response 12: Thank you! The highlight of pie chart is to show the percentage of sections. The imprecise comes from the unmarked numbers. Now the percentage of each section is added on Figure 5 and 6. Precise words are also used when describing the figures.
Point 13: p.5 line 189 - evidence that demand for Pareto solutions is growing? A citation would support this claim.
Response 13: Thank you! Here the authors obtain this conclusion because more literatures attempt to generate a set of Pareto solutions for their IAM rather than a preferred solution. Now this sentence is rewritten to avoid the misunderstanding (see Section 3.3).
Point 14: pp.5 & 6 discussion of more complex Pareto solutions and the reversion to a single solutions suggests, to me, that they, in effect, resort to a SOO form of decision-making. If this is so then this should be commented on.
Response 14: Sorry for the confusion! Some optimization algorithms convert a MOO problem into a set of SOO problems and each SOO problem corresponds to a Pareto solution. It does not mean they change a MOO problem into a single SOO problem. Section 3.4 are reviewed to clarify the confusion.
Point 15: Citations for heuristic definition. While Silver et al. may have had a definition, Tversky and Kahneman's work on heuristics is much more seminal. It won Kahneman a Nobel Prize in Economics (Tversky's early death prevented his award).
Response 15: Thank you for the suggestion! Many scholars have defined heuristics. The authors read several Tversky and Kahneman's papers on heuristics. Their heuristics focus on the uncertainties and preference when making decisions. They won Nobel Prize for their work on psychology of decision-making and behavioural economics. The authors think this manuscript needs a more general definition of heuristics.
Point 16: Classification of the 3 types of Applied Optimization Methods. I am not convinced that the software classification is a sufficiently distinct category to warrant its treatment as such. To my mind, the software algorithms uses variants of deterministic and/or heuristic methods. This suggests that much better categorisation and classification justification is required. I'd recommend Philip M. Podsakoff, Scott B. MacKenzie, and Nathan P. Podsakoff (2016) 'Recommendations for Creating Better Concept Definitions in the Organizational, Behavioral, and Social Sciences', Organizational Research Methods. Vol. 19(2), 159-203 as a useful reference in achieving greater conceptual definition. The issue is also relevant to p.6 lines 235-238.
Response 16: Thank you for the suggestion! Some reviewed literatures solve their problems by well-designed software tool. The authors cannot clearly classify these software tool into deterministic methods or heuristics because:
(1) These literatures do not clearly introduce the algorithms of the software tools.
(2) Some software tools may involve hybrid algorithms or several algorithms, which cannot be simply classified into deterministic methods or heuristics.
Hence, the manuscript leaves these literatures as the third class. Explanation is added in Section 3.4.
Point 17: Figure 5 has similar issues to those denoted above for earlier figures. It is unclear what the inner and outer rings denote. This is compounded by black and white reproduction issues but there does not appear to be supporting textual description either. Given previous comments about pie charts I suggest that a table would be more precise. The sequence of the discussion in the text does not follow that implied in the order of the legend items. For instance, Enumeration is first in the legend but appears halfway through the section.
Response 17: Thank you! Figure 6 (original Figure 5) is simplified and percentage of each pie section is added as well as the annotation.
The description does not follow the order of legend but the popularity of the algorithms. The algorithms with higher percentage, i.e. more frequently applied, are introduced earlier.
Point 18: p.6 line 227 expert adjustment to methods does raise all sorts of conceptual issues because while there is an underlying assumption that experts make technically rational decisions. It is just as likely that they use things like expert heuristics (rules of thumb learnt over the course of acquiring their expertise) and possible even (shock, horror!) subjective evaluations.
Response 18: Yes, expert adjustment is an important input. Here the authors do not mean expert adjustment is bad. If the algorithm requires expert adjustment, then knowledge is needed to support the expert adjustment before or during optimization procedure. On the other side, if the algorithm does not require expert adjustment during optimization, then experts can contribute based on the mathematical optimization result.
This paragraph in Section 3.4.1 is re-written to avoid confusion.
Point 19: Check the line of argument to ensure that consistent claims are made. pp.7 & 8 have at least 3 that I can identify. These should be clarified.
line 272 contradicts earlier statements in line 257.
line 288 appears to contradict line 287.
lines 298 & 299 contradicts line 292
Response 19: Thank you for pointing these out! They are not contradiction, but poor statement. These statements are re-written.
Point 20: p. 8 line 313 Please clarify what is meant by 'financial expense of management'. Is it the expense of the management itself, or is it, more likely from my experience, the financial expense of M&R?
Response 20: Thank you! Here the expense refers the cost of all the designed management interventions, not only maintenance and rehabilitation. It is reviewed.
Point 21: lines 316-317 suggest something else again, perhaps, a broader categorisation of 'expenses' though here imprecision is evident. Are 'users' the users of decision-making methods or infrastructure users? Is 'treatment cost' referring to M&R treatments or something else? Does 'agency cost' equate to costs of the (government, usually) agency responsible for the asset management or does it, invoking agency theory refer to a different form of agency?
Response 21: Thank you! Literatures differently define “expenses”, hence a precise “expenses” cannot be obtained from these literatures. However, clarification is added.
(1) The “users” are the infrastructure users.
(2) The “treatment cost” refers to M&R treatments
(3) The “agency cost” refers to the cost of asset management agency.
These clarifications are added in the manuscript.
Point 22: The last couple of outcome paragraphs - on level of service and other outcomes - appear to be afterthoughts, particularly the last one. Their significance is non commented on, merely listed.
Response 22: Thank you! These outcomes are not afterthoughts. Literatures consider different outcomes, and a wide range of outcomes are involved. This manuscript only introduces the details of the most popular ones, and other outcomes are only briefly mentioned. Comments on the outcomes are added in Section 3.5.
Point 23: Conclusion, last paragraph. Given the objective of the paper and the commentary about some of the challenges in application, particularly around complexity and scale of decision-making it seems useful to include some consideration of these things in the concluding statements.
Response 23: Thank you! Some thinking was added in the last paragraph of the manuscript.
Reviewer 2 Report
Please check the attached file for my comments.
Comments for author File: 
Comments.pdf
Author Response
Dear Reviwer,
Thank you very much for your comments! They are very valuable and helpful. Now the manuscript has been reviewed according to the comments. Please find the attachment for the responds to all the comments!
Thank you again!
Regards,
Lin
Point 1: I liked the idea of the paper of having a review for the critical element in decision making, which is optimization, regards infrastructure assets. However, there are some important points are needed to be clarified.
The paper was built mainly depending on literature review. However, the authors did not
Provide clear information and explanation for the following points.
Response 1: Thank you very much for the comment! Clarifications are made and the manuscript are reviewed.
Point 2: The papers’ selection. What are the procedure that the authors followed in selection and filtration of the papers?
Response 2: Thank you! The papers were found from main literature databases according to specific key words. Explanation is added in the manuscript in Section 3.
Point 3: There is no clear explanation about when asset management has started. This point is very important to show the evolution in infrastructure management systems though last decades
Response 3: Thank you! There is not a very clear starting point of asset management, while some important attempts during 70s, 80s and 90s are added and introduced in Section 3.1.
Point 4: The above clarifications will increase the reliability of the conclusions and suggestions that the paper provides. Also, the readers will get better understanding about different management systems from project management focusing systems in 70s to asset Management system nowadays.
Response 4: Thank you for the comment! The manuscript is reviewed and the main attempts in 70s, 80s and 90s are added.
Point 5: Moreover, the authors used infrastructure asset management title to all management systems such as pavement management systems, which was started in 70’s infrastructure asset management. However, infrastructure asset management systems have been used in practice and academia in late 90’s. They should be more clear about that.
Response 5: Thank you! Infrastructure Asset Management was applied in 70’s and references were provided to support the statement. The system was a relatively new concept and was adopted by Arizona Department of Transportation in 1980s. The supporting reference was also provided. Then the infrastructure asset management systems were widely admitted in 1990s. These are added in Section 3.1.
Point 6: Also, they should provide a clear definition for the infrastructure asset to make sure that all readers are on the same page.
Response 6: Thank you for the comment! Different organization has different definition of Infrastructure Asset Management. Several lines are added to introduce the definition of Infrastructure Asset Management in Section 1.
Point 7: There are no many mistakes in the manuscript. Based on the changes that will be done in figures, the manuscript will need to be changed to include numbers in the discussion.
Response 7: Thank you, the figures and the corresponding parts are updated.
Point 8: Figure 1 presents SOO and MOO, where both have not been mentioned earlier and they are not clear to reader what the authors do mean by them until section (3.3) in the next page.
Response 8: Thank you for the comment! Now Figure 1 was re-drawn and split into two individual figures, Figure 1 in Section 3.1 and Figure 4 in Section 3.3.
Point 9: Both Figure 1 and 2 present the number of literature that were used in this study but the procedure for the literature selection is not clear to reader in the manuscript.
Response 9: Thank you for the comment! The procedure for the literature selection is added in Section 3.
Point 10: Figure 4 doesn’t show any number for the literature in each section.
Response 10: Thank you for the comment! The number is added in the figure.
Point 11: Figure 5 doesn’t show any number for each optimization method also it’s difficult to be captured. Combining both optimization types and optimization methods makes it difficult to get any clear conclusion
Response 11: Thank you for the comment! The number is added on the figure, and Figure 5 (now Figure 6) was simplified.
Round 2
Reviewer 1 Report
Thank you to responding to the matters raised in a previous review. I still think that there are a couple of relatively minor matters that require attention before acceptance. These are:
Figure 1's abbreviations - MOO & SOO - still require a full definition in the legend for Figure 1.
Categorisation in Sections 3.4.1 to 3.4.3, particularly the last one. I am still not yet completely convinced that Software, as such, exists as a discrete category comparable to the deterministic and heuristic methods. Software is, these days, pretty much necessary to use either of these methods. What I think is the main point being made in Section 3.4.3 is that there is software that uses multiple methods or algorithms. Perhaps the section can be re-titled something like 'Multi-method software' with some adjustment to the text that recognises the the use of software for deterministic and heuristic methods, and that there is another category of software that employs multiple methods which is then described?.
Author Response
Dear Reviewer,
Thank you for your comments! The paper is improved accordingly. My respond to your comments is attached.
Thanks again for your time!
Regards,
Lin
Point 1: Thank you to responding to the matters raised in a previous review. I still think that there are a couple of relatively minor matters that require attention before acceptance.
Response 1: Thank you for all the comments! The manuscript is improved accordingly.
Point 2: Figure 1's abbreviations - MOO & SOO - still require a full definition in the legend for Figure 1.
Response 2: Thank you for the suggestion! Figure 1 (now Figure 4) is improved and the legend does not use abbreviation but the full definition.
Point 3: Categorisation in Sections 3.4.1 to 3.4.3, particularly the last one. I am still not yet completely convinced that Software, as such, exists as a discrete category comparable to the deterministic and heuristic methods. Software is, these days, pretty much necessary to use either of these methods. What I think is the main point being made in Section 3.4.3 is that there is software that uses multiple methods or algorithms. Perhaps the section can be re-titled something like 'Multi-method software' with some adjustment to the text that recognises the the use of software for deterministic and heuristic methods, and that there is another category of software that employs multiple methods which is then described?
Response 3: Thank you for the comment! The authors totally agree your comments on the importance of the Infrastructure Asset Management software. However, the software tools, especially the commercial ones, do not disclose their detailed algorithms. If the algorithm of the software is clearly introduced, this is classified into its corresponding class of deterministic methods or heuristics. Here the third class only covers the software tools that do not have disclosed algorithms. The explanation is added in Section 3.4.
