Differential Search Algorithm for Multiobjective Problems

In this paper, a novel Differential Search Algorithm (DSA) approach is proposed to solve multiobjective optimization problems, called Multiobjective Differential Search Algorithm (MODSA). MODSA utilizes the concept of Pareto dominance to determine the direction of a super-organism and it maintains non-dominated solutions in the external repository. This approach also uses the external repository of super-organisms that is used to guide other super-organisms. It guides the artificial organisms to search towards non-crowding and external regions of Pareto front. The performance of proposed approach is evaluated against the other well-known multiobjective optimization algorithms over a set of multiobjective benchmark test functions. Experimental results reveal that the MODSA outperforms the other competitive algorithms for benchmark test functions.