Inherited competitive swarm optimizer for large-scale optimization problems

Abstract

In this paper, a new Inherited Competitive Swarm Optimizer (ICSO) is proposed for solving large-scale global optimization (LSGO) problems. The algorithm is basically motivated by both the human learning principles and the mechanism of competitive swarm optimizer (CSO). In human learning principle, characters pass on from parents to the offspring due to the ‘process of inheritance’. This concept of inheritance is integrated with CSO for faster convergence where the particles in the swarm undergo through a tri-competitive mechanism based on their fitness differences. The particles are thus divided into three groups namely winner, superior loser, and inferior loser group. In each instances, the particles in the loser group are guided by the winner particles in a cascade manner. The performance of ICSO has been tested over CEC2008 benchmark problems. The statistical analysis of the empirical results confirms the superiority of ICSO over many state-of-the-art algorithms including the basic CSO.