Loss Model Approximations and Sensitivity Computations for Wireless Network Design

We develop and evaluate a new method for estimating various performance metrics of multi-hop wireless networks, including MANETs. The estimates depend implicitly on network design parameters, and therefore the method can form the basis for designing such networks so as to achieve predictable performance bounds. Explicit simulation models for large wireless networks are too complex and can not be used for analysis and design of wireless networks in realistic settings. We introduce a simple approximate (throughput) loss model that couples the physical, MAC and routing layers. The model provides quantitative statistical relations between the loss parameters used to characterize multiuser interference and physical path conditions on the one hand and traffic rates between origin-destination pairs on the other. We use a fixed point approach together with recently proposed models for the MAC layer to find a consistent solution satisfying all equations in the set. The result is an implicit model of performance metrics, such as throughput, packet loss, delay, parameterized by design variables, such as path routing probabilities, power levels, noise levels. We next apply Automatic Differentiation (AD) to these implicit performance models, and develop a methodology for design and parameter selection/tuning of wireless protocols.