enAHA: Enhanced Artificial Hummingbird Algorithm

Performance of the developed enAHA is validated on global optimization (CEC 2020 and CEC 2022) and three engineering design problems

Enhanced artificial hummingbird algorithm for global optimization and engineering design problems

The artificial hummingbird algorithm (AHA) is a recently introduced versatile metaheuristic optimizer that simulates flight patterns and intelligent foraging skills of hummingbirds. It has gained widespread recognition for its simplicity and adaptability to a wide range of optimization problems. However, the limited ability of the algorithm to establish the exploration-exploitation balance leads to getting stuck in local solution traps and premature convergence. To eliminate these drawbacks, this study introduces an enhanced artificial hummingbird algorithm (enAHA) based on a dynamic fitness-distance balance (dFDB) strategy. dFDB offers the opportunity to precisely balance exploration and exploitation throughout the optimization process with its dynamically adjustable weight coefficient. The convergence rate of the developed enAHA is tested on CEC 2020 and CEC 2022 benchmark problems. The enAHA and the original AHA results are statistically analyzed with the Wilcoxon signed-rank test. As per Wilcoxon test results, the proposed enAHA outperforms the original AHA algorithm for 70 %, 50 %, and 70 % of the CEC 2020 problems in 30-, 50-, and 100-dimensional optimization, respectively. In the CEC 2022 test suite, the enAHA showed a success rate of 58.33 % and 91.66 % with 10- and 20-dimensions. Moreover, the optimization capacity of enAHA is compared with the 29 state-of-the-art optimizers using the Friedman-rank test. Accordingly, the proposed enAHA algorithm ranked 1st, while the original AHA ranked 9th among the 30 competing algorithms. The practicability of the enAHA is validated on three engineering design problems: i) single-diode solar cell (SDSC) parameter extraction, ii) double-diode solar cell (DDSC) parameter estimation, and iii) optimization of pressure vessel design. The developed method provided minimum RMSE values of 7.730064E-04 for the SDSC and 7.422194E-04 for the DDSC. The enAHA algorithm achieved the best cost with a value of 5885.332773 for the pressure vessel design problem. Given that all experimental results are together, it is observed that the proposed enAHA algorithm can explore search space more efficiently and find competitive solutions compared to the original AHA and other compared ones.