An Improvised Competitive Swarm Optimizer for Large-Scale Optimization

In this paper, an improvised competitive swarm optimizer (ICSO) is introduced for large-scale global optimization (LSGO) problems. The algorithm is fundamentally inspired by the competitive swarm optimizer (CSO) algorithm which neither remembers the personal best position nor global best position to update the particles. In CSO, a pair-wise competition mechanism was introduced, where the particle that loses the competition is updated by learning from the winner and the winner particles are simply passed to the next generation. The proposed algorithm introduces a new tri-competitive mechanism strategy to improve the solution quality. The algorithm has been performed on different dimensions of CEC2008 benchmark problems. The empirical results and analysis have shown better overall performance for the proposed ICSO than the CSO and many state-of-the-art meta-heuristic algorithms.