A Novel Dynamic Neighborhood Learning Enhanced Artificial Gorilla Troops Optimizer for Global Optimization

The recently proposed artificial gorilla troops optimizer (GTO) has achieved satisfactory optimization outcomes. However, the canonical GTO performs exploration and exploitation in each iteration, leading to a loss of population diversity, premature convergence and trapping in local optima. To address these issues, this paper proposes a dynamic Neighborhood learning enhanced artificial gorilla troops optimizer (NGTO), incorporating a novel dynamic neighborhood learning (DNL) mechanism into the GTO framework. The DNL strategy guides individuals to learn dimensionally from dominant individuals in their immediate neighborhoods, improving regional search and population diversity. In the experimental section, the effectiveness of the DNL mechanism is first verified on the CEC 2017 benchmark suite, and NGTO is compared with nine other classical optimization algorithms. Out of 783 comparisons, NGTO wins significantly 732 times, draws 46 times and loses 5 times. In addition, optimization results on eight classical engineering design problems demonstrate that NGTO outperforms other algorithms and is superior in solving real-world problems. Furthermore, NGTO achieves the best optimization results in truss size optimization and obtains a new low truss mass of 389.1284 lb on a spatial 72-bar truss with discrete variables. In conclusion, the experimental results substantiate the exceptional capability of NGTO in addressing complex optimization challenges.