5G wireless networks will be extremely dense, given the projected increase in the number of users and access points (APs), as well as heterogeneous, given the different types of APs and the applications being accessed by the users. In such challenging environments, efficient mobility management (MM), and specifically user association, will be critical to assist the 5G networks in provisioning the Quality of Service (QoS) of diverse applications 5G targets to serve. Whilst determining the most suitable AP for the users, multiple constraints such as available backhaul capacity, link latency, etc., will need to be accommodated for. Hence, to provide an optimal user association solution, in this paper we present a joint optimization framework, namely AURA-5G. Under this framework we formulate our user association strategy as a Mixed Integer Linear Program (MILP) that aims to maximize the total sum rate of the network whilst optimizing the bandwidth assignment and access point selection. We analyze multiple active application profiles simultaneously, i.e. enhanced Mobile Broadband (eMBB) and massive Machine Type Communication (mMTC), in the network and study the performance of AURA-5G. Additionally, we provision a novel study on the multiple dual connectivity modes, wherein the user can be connected to either one macro cell and a possible small cell, or with any two favorable candidate access points. Utilizing the AURA-5G framework, we perform a novel comparative study of all the considered scenarios on the basis of total network throughput, performance against baseline scenario and system fairness. We show that the AURA-5G optimal solutions improve the performance of different network scenarios in terms of total network throughput (by 38$\times$–690$\times$) and system fairness as compared to the baseline scenario, which is a conventional user association solution. Further, we also present a fidelity analysis of the AURA-5G framework based on the backhaul utilization, latency compliance, convergence time distribution and solvability. And since, a given network cannot always guarantee to satisfy the future network loads and application constraints, we show that AURA-5G can be utilized by the operators/vendors to evaluate the myriad network re-dimensioning approaches for attaining a feasible and optimal solution. Henceforth, we then explore the possibility of network re-dimensioning and study its impact on system performance for scenarios where the performance of AURA-5G is severely impacted due to the extremely strict nature of the constraints imposed in the MILP.

鉴于用户和接入点（AP）的数量预计会增加，以及不同种类的AP和用户正在访问的应用程序，预计5G无线网络将非常密集。在这种充满挑战的环境中，有效的移动性管理（MM），尤其是用户关联，对于协助5G网络提供5G目标服务的各种应用的服务质量（QoS）至关重要。在确定最适合用户的AP的同时，将需要适应多个限制，例如可用的回传容量，链路等待时间等。因此，为了提供最佳的用户关联解决方案，在本文中，我们提出了一个联合优化框架，即AURA-5G。在此框架下，我们将用户关联策略表述为混合整数线性程序（MILP），旨在最大程度地提高网络的总和率，同时优化带宽分配和接入点选择。我们同时分析网络中的多个活动应用程序配置文件，即增强型移动宽带（eMBB）和大规模机器类型通信（mMTC），并研究AURA-5G的性能。此外，我们对多种双连接模式进行了新颖的研究，其中用户可以连接到一个宏小区和一个可能的小型小区，或者连接到任何两个有利的候选接入点。利用AURA-5G框架，我们基于总的网络吞吐量，相对于基准情景的性能和系统公平性，对所有考虑的情景进行了新颖的比较研究。$\times$–690$\times$）和与基线方案（这是常规的用户关联解决方案）相比的系统公平性。此外，我们还基于回程利用率，等待时间合规性，收敛时间分布和可解性对AURA-5G框架进行了保真度分析。而且，由于给定的网络无法始终保证满足未来的网络负载和应用限制，因此我们证明，运营商/供应商可以利用AURA-5G评估无数的网络重新规划方法，以获得可行和最佳的解决方案。从此以后，我们将探讨网络重新标注尺寸的可能性，并针对由于MILP施加的约束的极其严格的性质而严重影响AURA-5G的性能的情况，研究其对系统性能的影响。