kumoh national institute of technology
Networked Systems Lab.

Review Comment

NSL > Works@NSL> About Review> Review Comment
Sanjay Bhardwaj and Dong-Seong Kim, " Dragonfly Perspective for URLLC in Industrial Wireless Network', ,IEEE Transction of Industrial Informatics, 2018
By :
Date : 2019-01-02
Views : 97

Reviewer: 1

Comments to the Author
Given that the only novelty in this paper is the application, the content relating to new information is not sufficient. Moreover, how the presented idea directly impacts radio communications in industrial environments has not been discussed. The dragonfly behaviour as it relates to latency and reliability should also be investigated from a communication theory approach. As this will allow evaluation of its reliability and latency potential from a communications perspective. While the SDI presented from an algorithm point of view shows that U-BDA outperforms ICA, PSO & GA, the paper hasnt presented how this concept relates to latency, delay, capacity, radio communication, packetization, communication protocols or network topology. These quantities are essential metrics for URLLC in itself.

Minor Comments
• Sections I, II, III & IV are quite similar to Ref 24. Sections II, III & IV can be put together to probably 2 sections at most with appropriate citation of Ref 24.
• Points 4 and 5 in Section III should reflect position of food/enemy.
• Figure 1A is poorly presented.

Reviewer: 2

Comments to the Author
the idea of using biological dragonfly evolutionary algorithms in the context of URLLC is interesting. The use case of interest is that of industrial control with stringent constraints.
Having said that, the article does not provide much insights into the mathematical formulation of the system design. Section III is not enough but serves a background review on known literature. Section IV is too short. The other sections likewise are way too short.
Based on this , i do not recommend this article to be accepted. Lots of efforts are needed in having it in a proper shape to be reviewed.

Reviewer: 3

Comments to the Author
The paper investigates Ultra-Biological Dragonfly Algorithm (U-BDA) and relevant applications in industrial systems, authors are requested to:

-Add new figure at the beginning of the paper to summarize the relevant applications that their studied algorithm may support in industrial world.
-It is not clear what cloud be the improvements that this algorithm may add as authors turn the whole focus on analyzing the algorithm rather than use cases.
-References need to be significantly improved to tackle new literature.

Reviewer: 4

Comments to the Author
The paper suggests the use of a novel optimization technique, called ultra-biological dragonfly algorithm (U-BDA), which takes inspiration from the behavior of biological dragonflies, to optimize ultra-reliable and low-latency communication (URLLC) in industrial wireless networks.

Although the use of biologically-inspired optimization algorithms is interesting for configuring communication networks, I believe this paper fails to make a significant contribution in this field and hence I do not recommend it for acceptance.

The main issue is that it is not clear at all which problems the authors are trying to solve in the URLLC framework. There is no clear system model (which wireless networks are being considered? what is the topology, traffic patterns, etc.?) and no clear object functions or constraints (how are latency and reliability defined?). The connection between swarms of dragonflies and URLLC industrial wireless networks is only vaguely mentioned in Sec. II, where the authors claim that "avoidance of collisions", "alignment" and "cohesion" are the characteristics of a URLLC communication system and that the system should focus on the "targets" while avoiding enemies such as "noise sources". I believe this explanation is totally unclear for people with industrial communication background. The paper should have a completely opposite structure, in my opinion: start from the description of the URLLC problem, extract a high level model and clearly explain how it can be mapped to the "swarm of dragonfly" scenario and only then discuss an optimization algorithm to solve the original problem. Furthermore, the performance analysis should be targeted at URLLC performance indicators (e.g., latency, reliability, number of nodes, etc.), rather than some vague "test functions".

Besides this major issue, I also believe that the language and organization of the paper (too many sections) can be improved a lot.

Reviewer: 5

Comments to the Author
Nowadays, several research are focused on emulate the biological behaviors or multiple species, but (in the reviewer opinion) it is not only to select a new specie (new pattern) to create some new application. The authors also have to prove the impact in the current proposal in the current Wireless technologies. It is real necessary to applied Dragonfly to optimize the Wireless network, and it is better than other technologies, a practical use case must be introduced in the new version.