Simanta Das

This research presents a Differential Evolution (DE)-optimized surface plasmon resonance (SPR) based biosensor for Dengue virus detection with improved performance compared to the conventional SPR design. The DE algorithm is employed to optimize the biosensor parameters, demonstrating excellent performance for dengue detection. The optimized design improves the angular sensitivity from $309.742^\circ/\mathrm{RIU}$ to $312.857^\circ/\mathrm{RIU}$, increases the detection accuracy from $0.477$ to $0.734\ \mathrm{deg}^{-1}$, and significantly enhances the Figure of Merit (FOM) from $87.34$ to $1013.545\ \mathrm{RIU}^{-1}$ and the quality factor (QF) from $114.33$ to $176.113$, while achieving a reduced Full Width at Half Maximum (FWHM) of $3.380^\circ$, demonstrating highly effective sensing performance. The sensor exhibits a strong linear relationship between resonance angle and refractive index ($R^2 > 0.93$) over the RI range of 1.29–1.42, confirming high precision and reliability.The proposed DE-optimized SPR biosensor enables accurate dengue diagnosis and shows strong potential for machine learning-driven biomedical decision systems and advanced point-of-care diagnostics. The BK7/Ag/TiO$_2$/BP SPR configuration demonstrates robust fabrication tolerance, maintaining stable sensing performance under $\pm(5\%$ - $10\%)$ variations in layer thickness. The optical response of the sensor is analyzed using the Transfer Matrix Method (TMM) and verified by Finite Element Method (FEM) simulations to ensure accuracy and reliability of the modeling. Further analysis with different plasmonic metals, dielectric coatings, and prism substrates confirms the robustness of the proposed structure. Overall, the proposed DE optimized SPR biosensor demonstrates the effectiveness of the Differential Evolution algorithm as a machine learning-based optimization method for superior dengue diagnosis.