This paper proposes to combine three different Differential Evolution (DE) variants viz. DE/rand/1/bin, DE/best/1/bin and DE/rand-to-best/1/bin in an island based distributed Differential Evolution (dDE) framework. The resulting novel dDEs with different DE variants in each islands have been tested on 13 high-dimensional benchmark problems (of dimensions 500 and 1000) to observe their performance efficacy as well as to investigate the potential of combining such complementary collection of search strategies in a distributed framework. Simulation results show that rand and rand-to-best strategy combination variants display superior performance over rand, best, rand-to-best as well as best, rand-to-best combination variants. The rand and best strategy combinations displayed the poor performance. The simulation studies indicate a definite potential of combining complementary collection of search characteristics in an island based distributed framework to realize highly co-operative, efficient and robust distributed Differential Evolution variants capable of handling a wide variety of optimizations tasks.

Differential Evolution

Benchmark (surveying)

Source

Cite

Citations (0)

Taguchi method based parametric study of generalized generation gap genetic algorithm model

Lecture notes in computer science (2010)

S. Thangavelu C. Shunmuga Velayutham

Parametric model

Source

Cite

Citations (0)

Evolving malware variants as antigens for antivirus systems

Expert Systems with Applications (2023)

Ritwik Murali T. Palanisamy C. Shunmuga Velayutham

Cryptovirology

Jaccard index

Code (set theory)

Similarity (geometry)

10.1016/j.eswa.2023.120092

Cite

Citations (6)

Rough set based unsupervised feature selection in digital mammogram image using entropy measure

K. Thangavel C. Shunmuga Velayutham

Feature selection (FS) has become one of the most active research topics in the area of data mining. It performs to remove redundant and noisy features from high-dimensional data sets. A good feature selection has several advantages for a learning algorithm such as reducing computational cost, increasing its classification accuracy and improving result comprehensibility. In the supervised FS methods various feature subsets are evaluated using an evaluation function or metric to select only those features which are related to the decision classes of the data under consideration. However, for many data mining applications, decision class labels are often unknown or incomplete, thus indicating the significance of unsupervised feature selection. However, in unsupervised learning, decision class labels are not provided. The problem is that not all features are important, since some of the features may be redundant, and others may be irrelevant and noisy. In this paper, a novel unsupervised feature selection method using rough set based entropy measures is proposed. A typical mammogram image processing system generally consists of image acquisition, pre-processing, segmentation, feature extraction and selection, and classification. The proposed unsupervised feature selection method is compared with different supervised feature selection methods and evaluated with fuzzy c-means clustering inorder to prove the efficiency in the domain of mammogram image classification.

Contextual image classification

Supervised Learning

Feature (linguistics)

10.1109/icobe.2012.6178946

Cite

Citations (3)

Neural Network Based Image Registration Using Synthetic Reference Image Rotation

Lecture notes in computational vision and biomechanics (2019)

S. Phandi C. Shunmuga Velayutham

Image registration

10.1007/978-3-030-00665-5_88

Cite

Citations (1)

A preliminary investigation into automatically evolving computer viruses using evolutionary algorithms

Journal of Intelligent & Fuzzy Systems (2020)

Ritwik Murali C. Shunmuga Velayutham

This paper attempts to employ Evolutionary Algorithm(EA) techniques to evolve variants of a computer virus(Timid) that successfully evades popular antivirus scanners. Generating authentic variants of a specific malware results in a valid database of malware variants, which is sought by anti-malwar e scanners, so as to identify the variants before they are released by malware developers. This preliminary investigation applies EAs to mutate the Timid virus with a simple code evasion strategy, i.e., insertion and deletion(if available) of a specific assembly code instruction directly into the virus source code. Starting with a database of over 60 popular antivirus scanners, this EA based approach for malware variant generation successfully evolves Timid variants that evade more than 97% of the antivirus scanners. The results from these preliminary investigations demonstrate the potential for EA based malware generation and also opens up avenues for further analysis.

Code (set theory)

Evasion (ethics)

Computer virus

10.3233/jifs-179732

Cite

Citations (5)

A Comparative Study on Parameter Estimation of COVID Epidemiological Models Using Differential Evolution Algorithm

Studies in computational intelligence (2022)

Sai Sudha Panigrahi Arul Jayanth Muthukumar S. Thangavelu G. Jeyakumar C. Shunmuga Velayutham

Epidemic model

10.1007/978-981-16-8082-3_9

Cite