Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Сен. 19, 2024

Real-world optimization challenges frequently involve computationally expensive evaluations, necessitating efficient strategies. To address the demands of medium-scale problems, this research introduces and explores a novel Weighted Committee-Based Surrogate-Assisted Differential Evolution Framework (WCBDEF). This framework innovatively combines principles from active learning ensemble learning, iteratively interrogating most ambiguous high-fidelity solutions to ensure judicious allocation evaluation resources. WCBDEF employs dual sampling criterion, with offline dedicated exploration online focused on exploitation. Benchmarking against state-of-the-art surrogate algorithms across six test functions reveals that demonstrates clear advantage in performance, confirming its efficacy tackling optimization. Moreover, application optimizing operational parameters for two Enhanced Geothermal Systems (EGS) models has resulted significant reduction Levelized Cost Electricity (LCOE), surpassing existing algorithmic solutions. The results demonstrate significantly outperforms methods, exhibiting superior performance over single surrogate-assisted evolutionary (SAEAs) real-world thereby showcasing exceptional potential solving problems.

Язык: Английский