FSDE-Forced Strategy Differential Evolution used for data clustering
Meera Ramadas, Ajith Abraham, Sushil Kumar · Journal of King Saud University - Computer and Information Sciences · 2016
Differential evolution algorithm has seen various changes through numerous researches. Performance of the various algorithms depends on the changes in mutation and crossover strategies. Here in this paper, we are proposing a new variant of differential evolution named Forced Strategy Differential Evolution (FSDE), by creating a new mutation strategy. This strategy uses two parameters for mutation: a constant parameter and a variable parameter. FSDE will be applied on clustering using the k means technique. Experiments were conducted for various standard benchmark functions. FSDE was compared with the classical DE, GA and PSO in the field of clustering and the cluster quality results are tabulated. The results obtained show that the strategy implemented is more efficient than the other mutation strategies.