Optimal Resource Allocation in Wireless Networks for Multimedia Applications
Restricted (Penn State Only)
- Author:
- Wheatman, Kristina
- Graduate Program:
- Electrical Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- August 16, 2022
- Committee Members:
- Uday Shanbhag, Outside Unit & Field Member
Guohong Cao, Major Field Member
Vishal Monga, Major Field Member
Mark Mahon, Dissertation Co-Advisor
Thomas La Porta, Chair & Co-Dissertation Advisr
Thomas F La Porta, Program Head/Chair - Keywords:
- mobile networks
convex optimization
QoE
energy efficiency
video streaming
crowdsourcing
image processing
NUM
MINLP
objective function optimization
multimedia
utility
5G
LTE
base station
GPU
dual
primal
wireless
communications
markov chain
resource allocation
multi-path routing
mobile devices
cellular
network resource blocks
signal estimation
user experience
data download - Abstract:
- The prevalence of mobile devices has exponentially increased our ability to connect on an international level, swiftly accompanied by the popularity of communicating through the medium of multimedia. However, networking and processing requirements for exchanging images and videos are often orders of magnitude larger than those of text or even audio files. The research focuses on meeting and optimizing these pressing demands in the presence of competition for wireless network resources. Comprehensive solutions to energy-aware dual-path network utility maximization for crowdsourced image processing are derived and optimized. A delay resistant trace-based heuristic for distributed implementation that closely follows ideal algorithm behavior is developed. An iterative update solution to achieve maximum utility within a given time horizon using experimentally derived energy constraints is presented. Novel schemes for 5G network resource allocation and video streaming algorithms are provided. The research delivers a thorough analysis of all associated trade-offs in maximizing user Quality of Experience (QoE) while simultaneously minimizing associated energy consumption on mobile devices, as these goals are often in direct conflict with one another. Comparison of existing algorithms and proposed solutions are provided.