| Summary: | This thesis studies routing optimization problems arising in wired (Internet) and wireless network systems. Given network capacities and traffic demands, routing optimization is used to find a set of paths to forward packets from source to destination so as to optimize a well-defined performance function (such as packet delay, or link utilization). We first investigate an Internet routing optimization problem. In the Internet, the traffic demands change over time and are difficult to measure. We investigate the problem of finding a "robust" set of routes that results in good average case performance over changing or uncertain traffic demands, while avoiding poor worst case performance for such traffic demands. We then investigate the joint sensing and routing optimization problem in resource-constrained (such as energy-constrained) wireless sensor networks. In contrast to the Internet where the link capacity is relatively fixed, wireless networks arguably have more flexibility in adjusting their network capacity by changing their resource allocation strategies. Our goal is to determine the energy allocation between data sensing (gathering data) and data communication (transmitting this sensed data), and the data routes to maximize the aggregate amount of information received at the data sinks. Due to the need to respond to local unpredictable environmental changes (e.g., in the amount of energy needed to realize a given link capacity as a result of environmental changes), and the lack of centralized control of wireless sensor networks, we are especially interested in a distributed solution to this problem.
|
|---|
