DIAMOND: A Structured Coevolution Feature Optimization Method for LDDoS Detection in SDN-IoT

Table 1

Comparison with related works.

Reference	Technique	Advantage	Limitation
[7]	Whale optimization algorithm, binary variants	The two binary variants avoid a large number of local solutions.	More time and resources are consumed.
[8]	Binary grey wolf optimization	It is efficient in exploration and development.	This highly complex method potentially occupies more time and resource consuming.
[9]	Grey wolf optimizer algorithm, two-phase mutation	The two-stage mutation operator contributes to enhanced exploration.	More parameters need to be determined.
[10]	Hybrid whale optimization algorithm, simulated annealing	The algorithm reaches high accuracy by using a fewer number of features.	It has a high complexity and a tendency to fall into local optimal solutions.
[11]	Nondominated sorting algorithm adapted jumping gene operator	The jumping gene operator contributes to improve the detection accuracy.	Falling into local optimal solutions is a possible limitation.
[12]	The augmented whale optimization algorithm	It can search a more comprehensive range.	The convergence speed may be a little slow.
[13]	Multikernel SVM, gray wolf optimization	The multikernel SVM enhances the accuracy of the algorithm.	The multi-kernel SVM method may lead to overfitting.
DIAMOND	A structured coevolution feature optimization method for LDDoS detection in SDN-IoT	The model enables ordered individuals and coevolution of multiple subpopulations, reducing redundant computations.	How to coevolve subpopulations is an issue need to be addressed.