Exploring Delay Propagation Causality in Various Airport Networks with Attention-Weighted Recurrent Graph Convolution Method

Round 1

Reviewer 1 Report

Dear Authors,

 

I want to thank you for your patience, which was caused by a longer review process. Even so, I have to state that the article has a scientific character where the comprehensibility of the article is combined with the processed data. You evaluated these data in a good way, and with graphic-mathematical representation, you were able to create an idea of how it is possible to solve delays with the help of neural networks with a transparent comparison of airports.

The abstract of the article is comprehensible and identifies the essence of what I read in the article and verified with my judgment using the available software system.

At the beginning of the article, the authors created an overview in the form of a synthesis, in which they point to current scientific research on the given issue of different delays in the airspace-airport relationship.

Many causal problems related to delays also affect the technical equipment of the airport aimed at handling passengers, where slight problems, in the beginning, could be pointed out as a causal relationship with the management of the airport, which can cause delays. Such problems can also be caused by mechanisms that can be considered relevant in the given area.

According to the rules of citation, the introduction of the article is sufficient, after verification of the citations, the innovative current character of the given authors.

In Chapter 2. Methodology, the authors move precisely from a simple explanation of the applied neural network applied to the correlation of airports with delays to their simple proposal that clearly points to ongoing research.

Chapter 2.4 Attention Mechanism describes the problem of scales within airports. Is a functional model describable based on weight functions and transition functions?

In Chapter 2.5 Long Short-Term Memory Neural Network, a spatial block is used to present the delay time dependencies. Figure 6 illustratively solves this problem, which is not in the position of concrete implementation, but graphologically only outlines the possibilities with which the author develops thematically in the future. However, in the subsequent sequence of chapters, the reader gets lost, in what exactly was the need to illustrate Figure 6 with the subsequent algorithm.

Where already in chapter 2.7, the authors refer to the developed ARGC model. Was it introduced by the given algorithm from Chapter 2.6?

In the final chapter 4, it can be seen that the processing of flight routes with points representing airports, i.e. end and start points are understandable in Figure 9 and exactly readable in Table 9.

This is how the entire article continues, which shows a new way of exemplary situational management and comparison of selected airports in China with delays within the connected flight routes between them.

I must point out that the article has a scientific and informative value in the field of causal solutions to aircraft delays in flight operations. It is possible that the presented study could later be revised and completed with delays-related algorithms so that they can be used in future transport using multiple autonomous vehicles with the use of UAVs. That is low-flying aircraft (one-person) of the UAV type.

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

There is not much to tell on this very interesting manuscript, and just the following is suggested:

A) To what extent the presented methodology can be transferred elsewhere?  E.g. in Europe? Which are the basic requirements and/or caveats for networks of smaller size/operations? Please, elaborate

B) Figure 13a is barely readable, make it clearer and bigger

C) The conclusions look like a proxy for a short summary, whereas in this section limitations, potential, further steps of the resesrch work should be provided. Please, expand accordingly

Author Response

Please see the attachment.

Author Response File: Author Response.docx