Use of Optimization Models for Resource Allocation in Wireless Ad-Hoc and Sensor Networks

Optimization models and techniques are often used to achieve efficient allocation of limited network resources to competing demands in communication networks. In this talk, the speaker will give a brief overview of distributed optimization theory, including convex optimization problems for which iterative solution techniques exist and converge. The well-known Transport Control Protocol (TCP) is shown to be equivalent a distributed solution that achieves the optimal allocation of bandwidth in communication networks. As for wireless ad-hoc and sensor networks, each link capacity depends on the transmission power of other links due to co-channel interference. In addition, the quality of multimedia services supported by these networks cannot be represented by a concave function of the amount of allocated bandwidth. These factors unfortunately make the resource allocation problem for the wireless networks become a non-convex optimization problem. New distributed solution techniques will be presented to solve these problems and numerical examples will also be provided. This talk will also consider the in-network data processing in wireless sensor networks where data are aggregated (fused) along the way they are transferred toward the end user. It will be shown that finding the optimal solution for the distributed processing problem is NP-hard, but for specific parameter settings, the problem can lead to a distributed framework for achieving the optimal tradeoff between communications and computation costs. Future work on integrating data or signal processing techniques with the distributed solution framework will be discussed.