ISSN: 2966-0599

v.1, n.8, 2024 (Dezembro)

METADADOS

DOI: 10.69720/2966-0599.2024.0003

Author(1): Dauran

Biography(1): Department of Statistics, Usmanu Danfodiyo University, Sokoto, Nigeria.

Author(2): Aminu Bashiru

Biography(2): Department of Mathes & Statistics, Zamfara Collage of Arts and Science (ZACAS) Gusau,  Zamfara, Nigeria.

Author(3): Yakubu. Musa

Biography(3): Department of Statistics, Usmanu Danfodiyo University, Sokoto, Nigeria.

Author(4): Abdullahi A. Raji

Biography(4): Department of Vet. Patholoy, Usmanu Danfodiyo University, Sokoto, Nigeria.

ABSTRACT: The optimization of parameters to achieve desired performance outcomes presents significant challenges due to high dimensionality, collinearity among variables, and data noise. Traditional optimization methods, such as Response Surface Methodology (RSM), often face limitations when tackling these complexities. Additionally, the lack of efficient techniques for identifying and prioritizing influential parameters further exacerbates the difficulties in the optimization process. To address these intertwined issues, we propose an enhanced optimization methodology that integrates Principal Component Analysis (PCA) with Multi- Response Surface Methodology (MRSM). This integration aims to reduce dimensionality and highlight the most significant parameters, facilitating a more effective optimization process. We evaluated the effectiveness of the PCA-MRSM approach using a simulated dataset. Our findings indicate that the PCA-MRSM technique significantly outperforms traditional MRSM, leading to improved optimization outcomes and reduced variability. The analysis of response surface plots reveals a curved relationship between the principal components and the response variable, suggesting intricate interactions among the components. The results highlight the potential of the PCA-MRSM approach for optimizing complex systems characterized by multiple responses and correlated input variables. By effectively addressing the challenges of high dimensionality and collinearity, this methodology paves the way for more robust and insightful optimization techniques in intricate contexts.

Keywords: MRSM, PCA, DOE, Complex system, Optimization, Parameter Design