Utilizing the roulette wheel based social network search algorithm for substitution box construction and optimization
This paper introduces a new variant of a recent metaheuristic algorithm based on the Social Network Search algorithm (SNS), which is called the Roulette Wheel Social Network Search algorithm (SNS). As the name indicates, the main feature of RWSNS is the fact that the algorithm allows proportionate s...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English English |
| Published: |
Springer
2023
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/40732/1/Utilizing%20the%20roulette%20wheel%20based%20social%20network%20search%20algorithm.pdf http://umpir.ump.edu.my/id/eprint/40732/2/Utilizing%20the%20roulette%20wheel%20based%20social%20network%20search%20algorithm%20for%20substitution%20box%20construction%20and%20optimization.pdf |
| Summary: | This paper introduces a new variant of a recent metaheuristic algorithm based on the Social Network Search algorithm (SNS), which is called the Roulette Wheel Social Network Search algorithm (SNS). As the name indicates, the main feature of RWSNS is the fact that the algorithm allows proportionate selection of its search operators (i.e., from imitation, conversation, disputation and innovation) through exploiting the roulette wheel. Additionally, RWSNS also incorporates the Piecewise map as replacement for the pseudo random generator during the population initialisation to ensure high nonlinearity and allow further solution diversification. Finally, unlike its predecessor, RWSNS also permits the systematic manipulation of candidate solutions around the global best agent through the swap operator to boost its search intensification process, as the global best candidate solution is often clustered and always lurking around the current local best. Results based on the construction of 8 × 8 substitution-box demonstrate that the proposed RWSNS exceeds other competing metaheuristic algorithms in two main S-box criteria, namely, the average nonlinearity score and strict avalanche criteria (i.e., SAC offset), whilst maintaining a commendable performance on bits independence criteria, differential approximation probability and linear approximation probability. |
|---|