A hierarchical Pareto dominance based multi-objective approach for the optimization of gene regulatory network models

In this paper, a hierarchical Pareto dominance based multi-objective evolutionary approach is proposed for the optimization of gene regulatory network models. The approach is presented based on the neglected observations in GRN optimization that (i) structural dependencies exist among objectives; and (ii) some objectives may be more important than others. The hierarchical Pareto dominance is able to reduce the number of objectives during optimization process and increase the selection pressure to relieve the many objective problem. The proposed hierarchical Pareto dominance based multi-objective approach is verified and compared with classical Pareto dominance based algorithm NSGAII on the gene regulatory network optimization problem. The results obtained indicate that the presented approach has great performance when no noise exist. Also it shows superior results compared to NSGAII.