Optimization algorithm design for high-throughput wireless system
Title: Optimization algorithm design for high-throughput wireless system
DNr: SNIC 2022/22-1014
Project Type: SNIC Small Compute
Principal Investigator: Martin Andersson <martin.b.andersson@liu.se>
Affiliation: Linköpings universitet
Duration: 2022-10-18 – 2023-11-01
Classification: 20203
Keywords:

Abstract

The project aims to formulate optimization problems, design solutions, and to evaluate these. The optimization problems will aim to maximize the sum spectral efficiency for high-throughput wireless communication systems by jointly optimization several design parameters such as uplink power constraints, downlink power control, large-scale fading decoding weights and uplink/downlink resource allocations. Theoretical solutions to the formulated optimization problems will be designed and the computing resources will be used to evaluate the performance of these. The SNIC will be of use to speed up the process of solving different optimization problems and to evaluate several methods (e.g., successive convex approximation, accelerated projected gradient, alternating optimization and stochastic gradient decent) that are used to solve these optimization problems. Moreover, it will be needed to run a large amount of Monte Carlo simulations to get fair and comparable results of the throughputs of the different methods.