A Novel Approach for the Optimal Design of a Biosensor

Abstract

A novel design optimization strategy is proposed to enhance the analytical performance of a biosensor by taking into consideration the constructional and experimental parameters as design variables. A detailed study on multiple nonlinear neuro-regression analysis has been performed methodically in order to overcome the insufficient approaches on modeling-design-optimization of a biosensor. For this aim, the data were selected from a literature study. A hybrid method is used to test the accuracy of the predictions of 12 candidate functional structures that were proposed for modeling the data. The boundedness of the candidate models is checked after the calculation of R²_training and R²_testing values to reveal whether the model is realistic or not. Then appropriate models were optimized by using the four different optimization algorithms in terms of three different optimization scenarios. The results show that all the models express the process well regarding R²_training. However, only four models are appropriate based on R²_testing, and two of them were selected as the objective function depending on to be a realistic value. This novel optimization approach is also feasible for another modeling-design-optimization problem in analytical applications.

Keywords:

• Biosensor
• neuro-regression modeling
• optimization
• chemometrics