Advancing Network Utility Maximization for Intelligent Wireless Networks

Network Utility Maximization (NUM) serves as a theoretical control framework for resource allocation problems in modern communication systems, ranging from cellular downlinks to heterogeneous multi-hop wireless networks. NUM frameworks mainly seek to maximize global utility functions subject to system constraints arising from resource limitations (e.g., communication bandwidth, computing power, or energy availability). To optimally schedule these resources, classical NUM frameworks couple different network layers, utilizing feedback (e.g., queue length or channel state information) to solve the utility maximization problem at each slot. However, classical NUM renders inapplicable when utility functions are unknown, feedback from network layers are opaque or delayed, or when joint computing/communication resource scheduling are <a href="http://considered.

We” target=”_blank” title=”considered.

We”>considered.

We study the fundamental challenges in NUM frameworks involving multihop wireless networks and computing platforms, aiming to enable in-network computing and unified resource/service scheduling. Further, we investigate NUM in wireless networks characterized by unknown utility functions, partially observable routing actions, and channels/ queue states which are subject to non-trivial <a href="http://delays.

Speaker(s):” target=”_blank” title=”delays.

Speaker(s):”>delays.

Speaker(s): Mahfujul Kadir,

Virtual: https://events.vtools.ieee.org/m/565913