Computation offloading within Mobile Edge Computing (MEC) networks is a promising new technique, especially in the 5G era. This technique offers leading-edge services to the users of Smart Mobile Devices (SMDs) to reduce the processing time and battery drain. Thus, SMDs tend to offload their heavy processing tasks to preserve battery power and benefit from an important processing power. However, in the era of the Internet of Things (IoT), several subscribers will compete for the available provided resources. Thus, we consider subscribers with a priority property fixed by their contracts with the service provider. In this work, we study a multi-server MEC network with multiple base stations where each one is equipped with a MEC server and provides offloading services to nearby users. Accordingly, we consider the energy consumption, the critical situations of radio resources’ insufficiency as well as a penalty function based on the SMDs’ priority. Therefore, we formulated a bi-objective optimization problem that jointly minimizes the overall energy consumption and the penalty function while allocating the local processing frequencies for the SMDs, their transmit powers and the radio resources allocated by the Base Station (BS). Consequently, based on the weighted aggregation approach, we propose and study a heuristic solution called Resources Allocation with Priority Devices (RAPD). Finally, simulation experiments were realized to study the RAPD solution performance compared to some effective state of the art solutions, and the simulation results in terms of decision-making time, energy and penalty are very promising.

移动边缘计算（MEC）网络中的计算分流是一种很有前途的新技术，尤其是在5G时代。该技术为智能移动设备（SMD）的用户提供了领先的服务，以减少处理时间和电池消耗。因此，SMD倾向于减轻繁重的处理任务，以节省电池电量并从重要的处理能力中受益。但是，在物联网（IoT）时代，几个订户将争夺可用的提供的资源。因此，我们认为订户具有与服务提供商签订的合同所确定的优先级属性。在这项工作中，我们研究了具有多个基站的多服务器MEC网络，其中每个基站都配备了MEC服务器并为附近的用户提供卸载服务。因此，我们考虑了能源消耗，无线电资源不足的紧急情况以及基于SMD优先级的惩罚函数。因此，我们制定了一个双目标优化问题，该问题共同最小化了总体能耗和惩罚函数，同时为SMD，它们的发射功率和基站（BS）分配的无线电资源分配了本地处理频率。因此，基于加权聚合方法，我们提出并研究了一种启发式解决方案，称为优先设备资源分配（RAPD）。最后，通过仿真实验研究了RAPD解决方案性能，并与一些有效的现有解决方案进行了比较，并且在决策时间，精力和代价方面的仿真结果非常有希望。