Efficient, Sustainable and Intelligent Resource Allocation Approaches for the Next Generation of Wireless Communication Systems

点击翻译按钮获取中文摘要

The industrial and academic researchers are investigating the driving applications, enabling technologies and challenges of the future wireless communication systems (WCSs). Several pioneering and diverse applications and use-cases have been conceived, and they require stringent performance in terms of data rate, latency, reliability and massive access. Also, energy efficiency is seen as a major design criterion in future WCSs. The performance and energy efficiency gains expected from future WCSs can only be realized through challenging technological changes, such as reconfigurable intelligent surfaces, moving to higher frequency bands, user-centric cell-free massive multiple-input-multiple-output architecture, and drone-based wireless networks. Objectives - To allow the enabling technologies to achieve the stringent performance requirements of WCSs, the introduction of artificial intelligence (named "intelligentisation") is needed and seen as the major paradigm shift towards future WCSs. In particular, resource planning and management should exploit the progress in machine learning and knowledge sharing techniques to become more dynamic and cognitive in the future highly complex/dynamic WCSs. Accordingly, the main long-term objective of the Discovery Grant (DG) research program is to explore and propose innovative intelligent (learning-based), high-performance, and energy-efficient resource allocation and planning approaches and solutions for future WCSs. Our exhaustive literature survey allowed to specify several open problems to be solved for an efficient use of the emerging wireless technologies in future wireless systems and to define specific short-term objectives. Methodology - We will develop valid and tractable models of the wireless systems to be studied and formulate mathematical optimisation problems. Optimal and suboptimal algorithmic solutions will be designed based on centralized or distributed approaches. Next, we will focus on designing machine learning-based solutions for the formulated problems. In particular, we will extensively exploit the recent progress in reinforcement learning, federated learning and transfer learning to design our solutions. Impact - The benefits of this DG program are mainly in the areas of advancement of knowledge and industrial advances in WCSs. It will establish expertise, explore new research issues, and design efficient solutions towards the intelligentisation of energy-efficient WCSs. The theoretical and simulation results will contribute to both fundamental and practical knowledge, and provide exploitable insights and a better understanding of future WCSs. This DG program will contribute to the Canadian wireless industry by solving industrial problems and proposing practical intelligent algorithmic solutions for future WCSs. It will also allow a high-quality training of several students and thus meet the need of highly qualified people in several emerging areas of wireless industry.