Drones Routing with Stochastic Demand

Round 1

Reviewer 1 Report

This paper studies a multi drones routing problem with stochastic demand and proposes a heuristic algorithm to solve the problem. The investigated problem is interesting and valuable, but there are still some major changes required before the publication. The detailed comments are as follows: 

1) The challenges of solving the problem should be revisited and further explained. What is “two related constraints” in Line 77-78, and how are the two constraints interleaved with each other?

2) The literature review should be strengthened. A lot of studies about drone delivery has been published in the last two years.

3) The problem should be formulated more clearly, and some details such as problem assumption should be clarified.

4) In the four steps of the algorithm, it seems that the first three steps are a kind of classification considering a little robustness, while route planning in the fourth step is conducted after making customer demand clear. How is this different from those algorithms that directly plan the delivery route of multiple drones? Can you compare the solution quality of the proposed algorithm with such kind of algorithms?

5) The experiment mainly focuses on the advantage of the proposed algorithm in terms of cost saving. How to confirm the robustness under the stochastic demand?

 

The paper should be checked for grammatical errors and the label order of the references.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments to the authors:

1         It appears that the chapter on related works in the paper lists a large number of papers without adequately highlighting the differences or contributions of each paper to the research topic of Multiple Drone Collaborative Routing Design. Here are some suggestions for improvement:

(1)         Focus on relevance: While it's important to include relevant literature, it's crucial to ensure that the listed papers are directly related to the topic of the paper and its research questions.

(2)         Highlight differences and contributions: For each paper listed, provide a brief description of the differences or unique contributions of that paper to the research topic. This could include the methodology or limitations of the papers. This will help readers understand how each paper adds value to the overall research landscape and how it relates to the specific research being presented in the paper.

(3)         The Literature Review should be enhanced, especially on UAV/drone delivery and last mile delivery vehicle routing problem. These articles may be helpful for improving this paper: case study of drone delivery reliability for time-sensitive medical supplies with stochastic demand and meteorological conditions, the fourth-party logistics routing problem using ant colony system-improved grey wolf optimization, robust two echelon vehicle and drone routing for post disaster humanitarian operations, a hybrid metaheuristic algorithm for the multi-objective location-routing problem in the early post-disaster stage.

2         The omission of the takeoff and landing processes in the problem formulation of a Vehicle Routing Problem with Drones (VRPD) can lead to several limitations in the research. Here are some potential shortcomings:

(1)         Incomplete Comparison with Truck-based Routing: If the takeoff and landing processes are not considered in the problem formulation, the results of the research may be more akin to truck-based routing rather than providing a comprehensive comparison between drone-based and truck-based routing. Takeoff and landing processes are fundamental differentiators between drones and trucks in terms of operational capabilities, constraints, and efficiency.

(2)          Lack of Realism: Ignoring the takeoff and landing processes in the problem formulation of a Vehicle Routing Problem with Drones (VRPD) can also lead to inaccurate assessments of the flying time of drones and may not be suitable for studying the problem with time windows. For example, if the takeoff process is not considered, drones may be assumed to be instantly available for deliveries, which may not be realistic. Similarly, if the landing process is not considered, drones may be assumed to be able to continue flying indefinitely, which may not be practical due to battery limitations or other operational constraints. This can result in solutions that do not comply with actual time window constraints and may not be suitable for studying time-sensitive routing scenarios accurately.

3         No assumptions.

4         The abbreviation MORO is not clear, not easy to understand.

5         There are problems with the full text language, for example, what we multiple in line 153, should be changed to what is multiple d by what.

6         The formulas should be numbered.

7         The definition of symbols should be summary in a table

8         Algorithm comparison only and your own proposed method for comparison, so that others will feel that you just choose a good method which you proposed, and do not know whether it is better than the traditional method.

9         Constraints have not been explained.

 

 

 

The language should be improved.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors proposed the routing scheme for drone deliveries in stochastic environments. They claimed that leveraging the advantages of drones could make deliveries more efficient. To prove their claim, the authors proposed a drone routing algorithm that considers the delivery destination, drone, and capacity of the drone, and demonstrated the effectiveness and superiority of the algorithm proposed by the authors through simulation.

 

However, the authors overlooked the basic characteristics of drones, such as the maximum takeoff weight (MTOW) and the impact of weight on the drone's maximum speed, maximum flight time and maximum distance. They also calculated the cost of their proposed routing scheme solely based on distance, not including energy efficiency. As mentioned earlier, for more efficient drone operations, the maximum takeoff weight and the weight of the freights being carried by the drone should be considered together. It appears that both the algorithm and evaluation proposed by the authors only consider distance and time window, and therefore, the direction proposed in this paper, while interesting and promising, needs to be re-designed with a consideration of the aspects related to drone flight characteristics.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I have no further questions.

Author Response

Thank you for your hard work.

Reviewer 3 Report

The authors mentioned that the authors ignored the impact of the weight in the drone delivery, which is crucrial factor to the delivery planning in every types of delivery including land, water, and air. Thus the model only consider about the distance and no other environmental factors.

Moreover, there are many routing problem solver including google's OR-Tools, which could solve multiple-vehicle routing problem within a resonable time and could consider variable constraints including travel distance limit, time window constraint, capacity(volume, weight) limit.

Therefore, the authors have to refine the cost model to make the model to be more realistic.

Author Response

Please see the attachment

Author Response File: Author Response.pdf