Box-Behnken Design for Optimization of Particle Swarm Optimizer for Artificial Neural Networks: Application to Lab-on-a-Disc Biosensors

This paper investigates a lab-on-disk biosensor with the aim of improving its performance by optimizing the particle swarm algorithm through the application of Box-Behnken Design (BBD). The study concluded that the optimal conditions for the Particle Swarm Optimization (PSO) parameters - social learning factor (c $_{1} =0.5$ ), cognitive acceleration factor (c $_{2} =2$ ), inertia weight (w =0.65), and swarm size (Ps =176) - resulted in a significant improvement in prediction accuracy, as evidenced by an R-squared value of 99.9% and a low RMSE of 0.05. The results demonstrate the exceptional effectiveness of Box-Behnken Design (BBD) in optimizing PSO parameters for Artificial Neural Networks (ANNs), resulting in improved performance of the lab-on-disk biosensor. These optimized conditions not only improve response time, but also hold potential for broader applications in microfluidic sensing technologies.