Diversity-enhanced adaptive golden jackal optimization based on multi-strategy and its engineering applications

Diversity-enhanced adaptive golden jackal optimization based on multi-strategy and its engineering applications

연구 분야: Analysis

학회: Cluster Computing

Nowadays, many real-world optimization problems are becoming increasingly complex, leading to a growing popularity of metaheuristics. Among various metaheuristics, the recently developed Golden Jackal Optimization (GJO) has attracted significant attention due to its flexibility and effectiveness in engineering. However, GJO suffers from poor diversity and limited exploration. To address these drawbacks, this paper proposes a multi-strategy improved algorithm called Diversity-enhanced Adaptive Golden Jackal Optimization (DAGJO). Jackals in DAGJO are assigned two novel roles with five enhanced search modes to increase solution diversity. Particularly, the elite diverse utilization strategy is introduced to fully leverage individuals with better fitness. Then, the multiple candidate mechanism and perturbation mechanism are employed to avoid premature convergence. Additionally, the alternating compound adaptive mechanism is designed to balance exploration and exploitation capabilities. To verify the effectiveness, 25 benchmark test functions and 12 CEC2022 functions are solved by DAGJO, along with 12 excellent algorithms including the latest GJO variants. Furthermore, to assess the performance of DAGJO in real-world scenarios, four constrained engineering design problems and a data-driven automotive crash safety optimization design case are also employed for experiments. The results indicate that DAGJO generally exhibits superior performance in both global optimization and engineering design problems.