Modelling and Optimization of Personalized Scenic Tourism Routes Based on Urgency

Round 1

Reviewer 1 Report

Regarding to this work, on the one hand, the content, the mathematics and the conclusions are correct, but on the other hand, this is an old topic, not novelty at all, and simpler that the way is analyzed here. Therefore, depending what you are looking for, you can publish this work (since is correct) or add some added value (for instance, machine learning).

 

The mathematics are correct, but there is not novelty at all. There are already many algorithms can be used to solve this problem.

This work proposes a personalized route planning model based on urgency, unlike traditional methods based in lowest cost. Actually is not correct sentence, since lowest cost is the target. The difference is what we call lowest call, for instance, number of hops, bandwitch, priority, delay, etc. In this paper, the lowest cost is focused in the urgency, justa n arbitrary paramiter that can be assigned to any router interface or link, for example.

This topic has already many solutions long time ago. For instance, we can modify Dijkstr algorithm, or some other with more paramters, like Cisco EIGRP, or just implement a simple algorithm saying when to apply this or othat interface/link/segment cost.

Regarding the methodology: Maybe machine learning based algorithm  instead of a static (or quasi static) algorithm.

The content, the mathematics, the conclusiones, are correct.

The references are appropriate.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper presents a personalized route planning model based on urgency, taking into account user preferences. An improved Genetic Algorithm is used to solve the problem and the computational experiments presented show the merits of the proposed algorithm compared to rival methods.

The paper is well-written and the problem and approach are presented and explained well. The results are also interesting and reasonable conclusions are drawn from them. Nevertheless, there are some concerns found throughout the paper, so the authors are kindly asked to address the following important comments before the paper can be accepted for publication.

1-    In the abstract POI is used without being previously defined.

2- Also in the abstract you refer to the “improved genetic algorithm”, improved than what and in what way? Please name the distinguishing features of your algorithm.

3-    Page 1, lines 29, 30 35&36: there is no need to use quotations for regular, short and undistinguished text (e.g., “fastest”, “most economical”).

4-  Page 1, line 42: what do you mean by AOP value (e.g. distance, cost, weight..)?

5-    Page 2, line 53: “two points for users”, do you mean “points of interest”?

6-   Page 2, the sentence in lines 93-95 is out of place. It should be put at the beginning of the section or removed because it has already been mentioned before.

7-    Page 3, line 103: what does LP+M stand for?

8-   Page 3, line 107: “travel cost between the interest point set and the interest point set” Do you mean “travel cost between the interest point set and another interest point set?

9-   Page 4, line 167: It is better to mention some details about this method here rather than just referring to [25], because this is an important point in the research (please see my comment also in Section 3.1 and 3.2)

10- Page 5, line 194, 195: Note the redundancy in the title and the first sentence in the definition.

11- Page 5 line 197: “…and planning a personalized scenic tourism route meeting the ??????? for user ?”; sentence is not correct. Do you mean (\Omega is the planning…)?

12- Page 5, line 209 “?_?,? = 1 represents whether the edge from node to node is included in the path; otherwise, the edges from node to node are not included in the path.” “whether” should be replaced with “that” here because you have already specified that ?_?,? = 1.  Also, “node to node” should be “node n_i to node n_j”.

13- In Figure 1: Input should be specified separately for each stage as explained in the text rather than combining all inputs in one box. In addition,  I-C Matrix, I-T Matrix and T-C Matrix have not been explained anywhere in the paper (please see my relevant comment #9).

14- At the end of section 3.1, you should mention that you will explain each stage separately in the coming subsections.

15- For Section 3.2, The title of this section is different than the corresponding stage in Figure 1. They should be consistent for clarity. In addition, you should divide this section into two subsections : user preference extraction and relationship between scenic spots modeling as also shown in Figure 1.

16- In Algorithm 1, It is not clear how strategies are combined for the selection of edges. For example, if edge e1 is chosen according to strategy 1, then do you choose the next edge e2 according to strategy 2? Or do you first choose all edges according to  strategy 1 and then choose again edges according to strategy 2?Also, does the order of strategies have any significance?

17- Algorithm1, line 19: it is not clear the purpose of this method.

18- The flowchart in Figure 2 is missing the integration of the resulting offspring with the new population, which should come after "offspring" and before testing the condition of "Termination Criteria".

19- Page 10, line 362: why use a different symbol of budget than the one used before (d instead of b)?

20- Page 11, line 378: what do you mean by effective edge?

21- Page 11, line 384 and Figure 3: It is better to show an example of how this happens. Also, it is better that the example applies the mutation to the same chromosome resulting from crossover showing a case of duplicate edges, and then how this is repaired.

22- Page 11, line 392: Do you mean “crossover” instead of “cross mutation”?

23- Algorithm 2, line 3: why do you use “or” not “and” here?

24- Algorithm 2, line 4: How do you determine which criteria to use here?

25- Also in Algorithm 2, The symbols used in this algorithm are confusing because they are very similar. it is better to use meaningful symbols. For example, instead of “chr” you can use “tempchromo”, instead of “chrcs” you can use “geneset”, instead of “chrcr” you can use “newchromo”, etc.

26- Page 13, “OSM platform” & “Ctrip and Mafengwo”: need links or references to these datasets.

27- Page 13, line 452: Note the missing reference. Also, why is this work not found in the related work? In fact, More details about the rival algorithms should be included in the related work section.

28-  Also in reference to the competitive algorithms, did you implement the benchmark algorithms yourself or you did you use their published results? Did these algorithms use the same datasets? If you implemented the algorithms yourself, did you tune their parameters for best performance? Please clarify this important point.

29- Page 14, line 491: “the memetic algorithm has weak local search ability and cannot fully explore edges with high scenic scores”: This statement is not convincing given that local search is an essential component of the memetic algorithm in general.

30- The color codes used in Figure 7 are different than the remaining figures. They should be consistent.

31- In the captions of Figures 8 and 9 “star” is used instead of “start”.

32- Are the results in Figures 10, 11, 12 &13 based on the best result obtained or the average results of each algorithm?

33- In reference to Table 3, It is known the GA like all other meta-heuristics. does not guarantee optimality. So using “Optimal Value” is not accurate. It should be “Best value”.

34- A major concern: in all your results you do not comment about the performance of the traditional GA compared to your algorithm! You only comment about the fastest algorithm and the memetic algorithm. It is very important to discuss how your “improved” algorithm compares to the traditional GA to show the advantages of the new parts that you have introduced.

35- Last but not least, all references after reference [25] are not cited in the text! Most of them are also irrelevant to your research. Unless, there is some mistake, this is unacceptable since you cannot include any reference unless you are referring to it in the text.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

No new suggestions.