Determination of ion exchange parameters by a neural network based on particle swarm optimization

Modeling the process of ion exchange in glass requires accurate knowledge of the self-diffusion coefficients of the incoming and outgoing ions. Furthermore, correlating the concentration profile of the incoming ions to a change in refractive index requires knowledge of the correlation coefficient. A novel method of a neural network based on a particle swarm optimization algorithm is considered. In the range of training, the performance parameters of ion-exchanged waveguides in any arbitrary experiment condition can be obtained easily and quickly. This method has the advantages of reliability, accuracy, and time efficiency, which are identified by simulation. Therefore, it has promise in both fields of investigation and applications. (C) 2008 Society of Photo-Optical Instrumentation Engineers.