A Review of Constrained Multi-Objective Evolutionary Algorithm-Based Unmanned Aerial Vehicle Mission Planning: Key Techniques and Challenges
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors1. In Subsection 2.3, why not consider real-time planning algorithms, such as deep reinforcement learning, which have strong adaptability and real-time performance? For example, Drones | Free Full-Text | Factored Multi-Agent Soft Actor-Critic for Cooperative Multi-Target Tracking of UAV Swarms (mdpi.com).
2. In figure.4, is it suitable to add trajectory optimization or trajectory generation, as trajectory planning usually first obtains the route points, and then trajectory optimization obtains the trajectory.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThis paper presents a comprehensive review and analysis of Unmanned Aerial Vehicle mission planning algorithms, including the results of the last twenty years of research on constrained multi-objective evolutionary algorithms. The advantages and disadvantages of the reviewed approaches are summarised, allowing the results of this review to be applied to the development of new mission planning tools, selecting optimal solutions according to specific requirements. The methodology of literature selection and analysis used can be used as an example for the education of young scientists, while the comparative analysis of the objectives and constraints of mission optimisation provides information on the relevance of their derivations in the contemporary context.
Among the shortcomings of the study, I would include the trivial future research directions - the identified integration into 6G networks only partially solves the existing connectivity problems, blockchain technology is difficult to reconcile with the limited computational capability and power supply of UAVs, and quantum computing is still a distant prospect. The proposed UAV performance metrics and algorithmic metrics to evaluate the algorithm's performance are also very general and lack quantification criteria.
Not all the derivatives are explained in formula (1). The A-algorithm (line 345) would be more accurately called the A* search algorithm, similarly in Figure 4.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper is very interesting. Product of good and hard work in the field, Because I work in the field too I have some questions:
1) How can we use those algorithms if we use commercial drones eg of DGI or other companies?
2) How can we use all this knowledge and information if we are planning the mission based on commercial software in which we can determine only the flight's altitude and how the drone will fly? ( to realize eg an orthophoto map)
3) The problems with weather etc normally are solved before the fly because if the weather does not allow the drone will not fly.
4) despite the work being scientifically impressive, I don't understand how I can make a relation with all this information and Knowledge with the practice. In addition, no technological information is given about which drones are used, which software etc
5) Last but not least I would like to propose an idea from the past: Exist some multicriteria analysis decision methods. Two of the most well-known at the past were ELECTRA AND PROMITHEE. It could be used in my opinion in the next phase of your research
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
