Wireless sensor networks (WSNs) have been pushed to the forefront in the past decade, owing to the advert of the Internet of Things. Our research suggests that the reliability and lifetime performance of a typical application in WSNs depends crucially on a set of parameters. In this paper, we implemented our experiments on the nearest closer protocol with the J-Sim simulation tool. We then analyze the closure relationships among the density, reliability and lifetime, and reveal the trade-off among them based on our analysis on the experiment results. Next, we propose five intelligent evaluation models that are applicable to such situations. Our research allows the WSN users to predict the significant evaluation parameters directly from the settings while costly simulations are no longer necessary.
