In the present work an in-depth investigation on direct and inverse modeling problems has been made using artificial neural network, fuzzy logic and evolutionary computing algorithms.In short, a low complexity artificial neural network has been used as back bone for achieving efficient identification and channel equalization tasks. The adaptive model has been created by training its internal parameters using DE, BFO, and PSO etc. In addition the work has embodied results pertaining to distributed identification for sensor network environment as well as robust identification particularly when outliers are present in the training samples itself.