Elitist Non-Dominated Sorting Directional Bat Algorithm (Ensdba)

A novel way of performing nondominated sorting in a multiobjective optimization problem is proposed using a modified directional Bat algorithm. Unlike NSGA-II, where the solutions of two generations are merged and then sorted for elitism, in the proposed algorithm, the solution is generated and compared with all the previous solutions one by one. Hence, this method reduces the computational time by avoiding the comparison of solutions of two generations, at the same time, generates a diverse solution. A unique way of sorting the solutions is proposed using a Nondomination matrix, which can easily be updated if a new solution is accepted. The Nondomination matrix serves as an archiving strategy to preserve elitism. Detailed criteria are proposed for the selection of a new solution. We have tested the proposed algorithm on some of the standard benchmark optimization problems. The results show that the proposed algorithm is very competitive and outperforms other algorithms in terms of efficiency and other performance metrics for most problems. The algorithm also provides a standard platform for nondomination sorting, which can be applied to any other metaheuristic algorithm.