Bald eagle search algorithm (BES) is a recent metaheuristic algorithm based on bald eagle hunting behavior. Like other metaheuristic algorithms, the BES algorithm is prone to entangle in local optimums due to limited diversity, sluggish convergence rate, or improper equilibrium between exploitation and exploration. Thus, adaptive parameters are injected into the original BES to overcome these shortcomings. These parameters are a function of the current and the max number of iterations. They provide the eagle with more diversity during the exploration and exploitation phases. The modified BES is tested on test functions provided by Congress on Evolutionary Computation 2020 and Congress on Evolutionary Computation 2022. The obtained results are compared to that of other reliable and recent algorithms. In addition, analysis of variance and Tuckey tests are utilized to confirm the results' significance. Due to its benefits, lithium-ion batteries are employed in more and more applications. However, extracting its parameters is challenging due to its complex model. Hence, the proposed algorithm will handle this task to approve its performance in complex problems. The significant benefit of this extraction method is its excellent precision, with fitness value declining (root mean square error) to 0.89 × 10-3 compared to the original BES (1.013 × 10-3) with a standard deviation of 1.12 × 10-3. To confirm the performance of mBES, a second battery was tested with the New European Driving Cycle profile. The mBES has the lowest fitness values (0.058896) and the least standard deviation (5.89 × 10-7).
Keywords: Bald eagle search algorithm (BES); Lithium-ion battery model; Metaheuristic optimization algorithms (MAs); Parameters identification.
Copyright © 2022 ISA. Published by Elsevier Ltd. All rights reserved.