Unlike 2G systems where the radius of macro base station (MBS) could reach several kilometers, the cell radius of LTE-Advanced and next generation wireless networks (NGWNs) such as 5G networks would be random and up to a few hundred meters in order to overcome the radio signal propagation impairments. Heterogeneous wireless networks (HetNets) are becoming an integral part of the NGWNs especially 5G networks, where small cell base stations (SBSs), wireless-fidelity (WiFi) access points (APs), cellular BSs and device-to-device (D2D) enabled links coexist together. HetNets represent novel approaches for the mobile data offloading, resource allocation and coverage probability problems that help to optimize the network traffic. However, heterogeneity and interworking among different radio access technologies bring new challenges such as bandwidth resource allocation, user/cell association, traffic offloading based on the user activity and coverage probability in HetNets. This dissertation attempts to address three key research areas: traffic offloading, bandwidth resource allocation and coverage probability problems in HetNets. In the first part of this dissertation, we derive the mathematical framework to calculate the required active user population factor (AUPF) of small cells based on the probabilistic traffic models. The number of total mobile users and number of active mobile users have different probabilistic distributions such as different combinations of Binomial and Poisson distributions. Furthermore, AUPF is utilized to investigate the downlink BS and backhaul power consumption of HetNets. In the second part, we investigate two different traffic offloading (TO) schemes (a) Path loss (PL) and (b) Signal-to-Interference ratio (SIR) based strategies. In this context, a comparative study on two techniques to offload the traffic from macrocell to small cell is studied. Additionally, the AUPF, small cell access scheme and traffic type are included into a PL based TO strategy to minimize the congested macrocell traffic. In the third part, the joint user assignment and bandwidth resource allocation problem is formulated as a mixed integer non-linear programming (MINLP). Due to its intractability and computational complexity, the MINLP problem is transformed into a convex optimization problem via a binary variable relaxation approach. Based on the mathematical analysis of the problem, a heuristic algorithm for joint user assignment and bandwidth allocation is presented. The proposed solution achieves a near optimal user assignment and bandwidth allocation at reduced computational complexity. Lastly, we investigate the transition between traditional hexagonal BS deployment to random BS placement in HetNets. Independent Poisson Point Processes (PPPs) are used to model the random locations of BSs. Lloyds algorithm is investigated for analyzing the coverage probability in a network which functions as a bridge between random and structural BS deployments. The link distance distribution is obtained by using the Expectation-Maximization (EM) algorithm which is further utilized for calculating the coverage probability.
Unlike 2G systems where the radius of macro base station (MBS) could reach several kilometers, the cell radius of LTE-Advanced and next generation wireless networks (NGWNs) such as 5G networks would be random and up to a few hundred meters in order to overcome the radio signal propagation impairments. Heterogeneous wireless networks (HetNets) are becoming an integral part of the NGWNs especially 5G networks, where small cell base stations (SBSs), wireless-fidelity (WiFi) access points (APs), cellular BSs and device-to-device (D2D) enabled links coexist together. HetNets represent novel approaches for the mobile data offloading, resource allocation and coverage probability problems that help to optimize the network traffic. However, heterogeneity and interworking among different radio access technologies bring new challenges such as bandwidth resource allocation, user/cell association, traffic offloading based on the user activity and coverage probability in HetNets. This dissertation attempts to address three key research areas: traffic offloading, bandwidth resource allocation and coverage probability problems in HetNets.
In the first part of this dissertation, we derive the mathematical framework to calculate the required active user population factor (AUPF) of small cells based on the probabilistic traffic models. The number of total mobile users and number of active mobile users have different probabilistic distributions such as different combinations of Binomial and Poisson distributions. Furthermore, AUPF is utilized to investigate the downlink BS and backhaul power consumption of HetNets.
In the second part, we investigate two different traffic offloading (TO) schemes (a) Path loss (PL) and (b) Signal-to-Interference ratio (SIR) based strategies. In this context, a comparative study on two techniques to offload the traffic from macrocell to small cell is studied. Additionally, the AUPF, small cell access scheme and traffic type are included into a PL based TO strategy to minimize the congested macrocell traffic.
In the third part, the joint user assignment and bandwidth resource allocation problem is formulated as a mixed integer non-linear programming (MINLP). Due to its intractability and computational complexity, the MINLP problem is transformed into a convex optimization problem via a binary variable relaxation approach. Based on the mathematical analysis of the problem, a heuristic algorithm for joint user assignment and bandwidth allocation is presented. The proposed solution achieves a near optimal user assignment and bandwidth allocation at reduced computational complexity.
Lastly, we investigate the transition between traditional hexagonal BS deployment to random BS placement in HetNets. Independent Poisson Point Processes (PPPs) are used to model the random locations of BSs. Lloyds algorithm is investigated for analyzing the coverage probability in a network which functions as a bridge between random and structural BS deployments. The link distance distribution is obtained by using the Expectation-Maximization (EM) algorithm which is further utilized for calculating the coverage probability.