Eight-Figure Pattern for Enhancing the Searching Process of Grey Wolf Optimization (Eight-GWO)

Evolutionary algorithms suffer significantly from a stack at the local optima. This paper proposes a new strategy that detects when the search gets stuck in a local optimum and then switches to a more dynamic approach to escape. The proposed model is based on simulating eight pattern movements and embedded with a Grey Wolf Optimizer algorithm (GWO). It is called the Eight-Figure Grey Wolf Optimizer (Eight-GWO). The proposed model combines two phases: regular search when searching progresses over time while the second phase, searching by eight patterns when the algorithm reaches stuck. The Eight-pattern updates the gray position based on the sin and cos function. The proposed Eight-GWO on the 24 functions of the CEC2005 benchmark suite and compared its results with both the standard GWO and Particle Swarm Optimization (PSO). The experiments result show the proposed Eight-GWO gets better results than GWO and PSO where it achieved the best results on 80% of the test functions. The proposed Eight-GWO runs 23% faster than the original GWO and 44% faster than PSO.