Adaptive Grey Wolf Optimizer for Global Numerical Optimization

As a well-regarded GWO variant, the improved Grey Wolf Optimizer (IGWO) is proven competitive in solving global numerical optimization problems due to its new proposed dimension learning-based hunting (DLH) search strategy. However, the high computational complexity of IGWO limits its potential application prospect. In this paper, a novel adaptive grey wolf optimizer with DLH (AGWO-D) is proposed to alleviate the heavy computational complexity of IGWO. The AGWO-D benefits from an execution probability model for each search agent to adopt the DLH search. By this mechanism, the search areas with more potential are more likely to be exploited, and thus the computational resources are used with higher efficiency. To evaluate the performance of the proposed AGWO-D algorithm, a simulation experiment is run on the CEC 2014 benchmark suite. And the experimental results show that, compared with IGWO, the AGWO-D algorithm has superior search capability with less computational time.