Design and comparison of discrete swarm intelligence algorithms for makespan minimization of parallel machines for complex RNA sequence allocation jobs

Abstract

This paper presents discrete combinatorial versions of popular swarm intelligence algorithms for makespan minimization on the systems connected through parallelization for the job distribution of highly complex non-identical biological sequence sets. For this purpose, 5000 RNA sequence sets are randomly generated, that varies from a minimum 5 and maximum 10,000 sequences in an individual sequence set with average length of 209 nucleotides with measured complexity 2.49E+14. For minimizing makespan time, discrete combinatorial versions of four swarm intelligence algorithms are designed and compared using statistical tests i.e. one-way ANOVA followed by post-hoc analysis by bonferroni test. The algorithms are evaluated at the basis of makespan completion time, algorithmic time to optimize makespan and convergence. The overall performance evaluation finds artificial bee colony algorithm as the best performer followed by spider monkey optimization algorithm.