On Gradient Descent Optimization in Diffusion-Advection Based 3-D Molecular Cooperative Communication.

This work considers a cooperative communication system in 3-D fluid medium in which the flow of molecules is supported by the drift and the diffusion phenomena. To enhance the system performance, the equal gain combining is used at the destination nanomachine (DN) where the molecular signals arriving from the direct and the cooperative paths are combined together by employing equal weights. Using the gradient descent algorithm, the optimum threshold at DN, and the optimal number of molecules transmitted from source and cooperative nanomachines are obtained. For this purpose, the convex constraints are determined using the closed-form expression for the average probability of error at DN. Finally, the accuracy of the analytical results is validated through the particle/ Monte Carlo-based simulations.