As online services and applications are moving towards a more human-centered design, many app vendors are taking the quality of experience (QoE) increasingly seriously. End-to-end latency is a key factor that determines the QoE experienced by users, especially for latency-sensitive applications such as online gaming, autonomous vehicles, critical warning systems and so on. Edge computing has then been introduced as an effort to reduce network latency. In a mobile edge computing system, edge servers are usually deployed at, or near cellular base stations, offering processing power and low network latency to users within their proximity. In this work, we tackle the edge user allocation (EUA) problem from the perspective of an app vendor, who needs to decide which edge servers to serve which users in a specific area. Also, the vendor must consider the various levels of quality of service (QoS) for its users. Each QoS level leads to a different QoE level. Thus, the app vendor also needs to decide the QoS level for each user so that the overall user experience is maximized. We first optimally solve this problem using Integer Linear Programming technique. Being an $\mathcal{NP}$-hard problem, it is intractable to solve it optimally in large-scale scenarios. Thus, we propose a heuristic approach that is able to effectively and efficiently find sub-optimal solutions to the QoE-aware EUA problem. We conduct a series of experiments on a real-world dataset to evaluate the performance of our approach against several state-of-the-art and baseline approaches.

随着在线服务和应用程序朝着以人为本的设计方向发展，许多应用程序供应商都越来越重视体验质量（QoE）。端到端延迟是确定用户体验的QoE的关键因素，尤其是对于延迟敏感的应用程序，例如在线游戏，自动驾驶汽车，紧急警报系统等。然后，为了减少网络等待时间而引入了边缘计算。在移动边缘计算系统中，边缘服务器通常部署在蜂窝基站或其附近，从而为附近的用户提供处理能力和较低的网络延迟。在这项工作中，我们从应用程序供应商的角度解决边缘用户分配（EUA）问题，后者需要确定哪些边缘服务器为特定区域中的哪些用户提供服务。也，供应商必须为其用户考虑各种级别的服务质量（QoS）。每个QoS级别导致一个不同的QoE级别。因此，应用程序供应商还需要为每个用户确定QoS级别，以使整体用户体验最大化。我们首先使用整数线性规划技术来最佳解决此问题。作为一个$\mathcal{NP}$-困难的问题，在大规模方案中以最佳方式解决它是棘手的。因此，我们提出了一种启发式方法，该方法能够有效且高效地找到QoE感知EUA问题的次优解决方案。我们在一个真实的数据集上进行了一系列实验，以评估我们的方法相对于几种最新方法和基准方法的性能。