Clustering big datasets using orthogonal gray Wolf optimizer

Grey wolf optimization (GWO) is one of the many nature-inspired algorithms that is popular for solving labyrinthine optimization problems. The algorithm uses properties of hunting mechanism and leadership hierarchy of Grey wolves. This manuscript proposes a new variant of GWO called as Orthogonal Grey Wolf Optimization (OGWO). It is different from the original GWO in a sense that the position of wolves are not merely updated by averaging the movement towards three global leaders. Instead a combination termed as orthogonal methodology is used to determine the effective update position of the leader wolves. Here the methodology objective is to obtain the best possible combination of positions from the three global leaders. The simulation analysis on standard benchmark function reveals that the results obtained from the proposed algorithm are more optimal and have lesser standard deviation than the previous approach. In addition to this, the proposed algorithm is also successfully used on cluster analysis and very competent results are obtained when compared to other nature-inspired algorithms like original GWO, Particle Swarm Optimization (PSO), Orthogonal PSO (OPSO), Orthogonal Genetic Algorithm with Quantization (OGA).