Improving the response surface methodology optimization with metaheuristics: A practical approach
Küçük Resim Yok
Tarih
2026
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Pergamon-Elsevier Science Ltd
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Response Surface Methodology (RSM) is a powerful statistical technique that explores the relationship between input and output response variables using statistical methods and mathematical tools throughout the process. Optimization in the RSM is carried out using deterministic models that usually fall into local optima, losing the possibility of finding parameters that achieve better results. To overcome this problem, this article proposes applying metaheuristic algorithms (MA) in the optimization phase of RMS. Here MA offers a solution by optimizing the models derived from RSM. Three real problems are used to verify the performance of the combination of MA with RSM: removal of chemical oxygen demand, biodiesel synthesis from Ceiba pentandra oil, and biodegradation yield of reactive blue. The experimental results validate that the RSM with MA optimize with better results than the models obtained from the analysis of variance. For experimental purposes nine MA were used: Covariance Matrix Adaptation Evolution Strategy (CMAES), Differential Evolution (DE), Estimation of Distribution Algorithm (EDA), Particle Swarm Optimization (PSO), Comprehensive Learning PSO (CLPSO), Runge Kutta Optimizer (RUN), Dung Beetle Optimizer (DBO) Liver Cancer Algorithm (LCA), and Sea-Horse Optimizer (SHO). The metaheuristic with the best results in solving the three problems was the DE. The DE obtained an improvement of 0.56% over the removal of the first problem, for the second problem it obtained the same result as the deterministic technique, but in the third problem that showed surfaces with greater complexity, an improvement of 5.92 % was obtained. This research is a starting point for work in which RSM is applied to solve real-world problems that generate complex models.
Açıklama
Anahtar Kelimeler
Response Surface Methodology, Metaheuristics, Optimization, Real-World Problems
Kaynak
Expert Systems with Applications
WoS Q Değeri
Q1
Scopus Q Değeri
Cilt
297
