Resource Allocation for 5G Systems

The 5th generation of cellular communications is driven by three broadly classified use cases, enhanced mobile broadband, massive machine type communication, and ultra-reliable low latency communication. These new use cases bring a new set of problems such as provisioning of extremely high user rates, ultra-high reliability, ultra-low latency, extremely low jitter, and extremely long battery life, to name a few. Therefore, in this thesis, we focus on lower protocol layers of radio access networks, designing technology components enabling communication confirming new extreme requirements introduced by 5th generation systems. To increase the amount of spectrum available for an operator in a specific geographic region, we discuss inter-operator spectrum sharing. The operators employ non-cooperative game-theoretic mechanisms to enable sharing without revealing their RAN-related information, e.g., optimization targets and traffic load to each other. The operators are independent and motivated by self-interest. However, if the sharing persists over a long time, non-cooperative operators may develop a degree of cooperation to maximize their self-interest by employing repeated games. Rational operators tend to cooperate due to fear of repercussions if they focus on optimizing long-term instead of short-term gains. Such cooperation emerges from individuals' greedy decisions. Simulation results show that the spectrum sharing based on repeated games provides better network utility than default spectrum allocations in the scenarios like limited spectrum pool or mutual renting. In contrast to spectrum allocation, which focuses on the rate improvement for most users in a mobile broadband scenario, the other half of the thesis pivots to providing extremely reliable service, which induces the need to concentrate on the extremely deficient users in a service area. To improve critical communication, we consider allocation designs that boost rate availability and reliability. The availability targets a specific rate that a cell can provide at probable locations. We inspect multi-hop mechanisms for extreme availability, focusing on time scheduling weights, power allocation, and precoding weights. Inter-cell interference can limit availability at the cell edge, and thus we consider coordination among cells in limiting the interference. Furthermore, we consider allocation designs for minimum rate improvement using transmission repetitions for low latency contention-based access services. Simulation results show that the employed techniques significantly improve 5G ultra-reliable, low latency services.