Boolean Binary Grey Wolf Optimizer

Resumo

Several binary swarm algorithms use the identical continuous proposal, adding a transfer function to mapping from continuous to binary space. It has been shown that binary operators are more appropriate and efficient for binary optimisation. Based on it, we proposed the Boolean Binary Grey Wolf Optimizer, a new version of Grey Wolf Optimizer to solve binary optimisation. Our proposal uniquely operates in the binary space using binary vector and boolean operators, precisely AND, OR, XOR, and NOT gates. We used OneMax, ZeroMax and 0-1 Knapsack to compare our proposal with bGWO1, bGWO2 and BPSO. In general, BBGWO outperformed the other swarm-based algorithms in different scenarios. Furthermore, analysing the swarm hamming distance and the unique candidate solutions through optimisation, we found that the proposal can overcome premature convergence, and most of the time, wolves are in different positions.

Publicação
2022 IEEE Latin American Conference on Computational Intelligence (LA-CCI)
Rodrigo Lira
Rodrigo Lira
Professor

Rodrigo Lira é professor no IFPE e tem interesse nas áreas de inteligência de enxames, aprendizado de máquina e IoT.

Próximo
Anterior