MOSDA: A Proposal for Multiple Objective Spiral Dynamics Algorithm

Abstract

This paper proposed a multi-objective spiral dynamic algorithm (MOSDA) to solve multiple objectives problems. SDA is originally a single objective optimizer that inspired based on the spiral phenomena in nature. It has a good elitism strategy and has a simple structure. A method called “archive method” that is used in multi-objective particle swarm optimization (MOPSO) is adopted into SDA to develop its multiobjective (MO) type algorithm. Moreover, MOSDA is formulated by applying the widely-used concept of Pareto dominance to determine the movement of the particles and at the same time, the algorithm maintains the non-dominated solution in a setup global repository. These non-dominated solutions then will be used to guide other particles to move. The proposed algorithm is tested with several benchmark functions for multi-objective problems. Pareto front (PF) graphs are presented as the results of these tests. The accuracy and diversity of the produced PF are highly competitive compared to MOPSO.