基于多目标优先级粒子群算法的资源调度策略

Resource Scheduling Strategy Based on Multi-objective Priority Particle Swarm Optimization

移动边缘云计算是5G技术的核心之一,也是当下非常热门的通信技术。但当前移动用户数量迅猛增长,传统资源分配方式已不能满足用户需求,因此根据用户的规模及其任务优先级的实时变化,如何合理制定资源分配策略来满足用户对计算单元、存储空间、软件等资源的需求是当下十分热门的研究方向。该文提出了一种基于多目标优先级粒子群算法的边缘云资源调度算法(MPPSO),合理布局多个边缘基站,形成边缘云。在多用户多任务并发时,综合用户数据传输速率、任务能耗、任务优先级和边缘基站性能等多方面因素,设计了两个适应度函数和一种粒子编解码方法,同时引入了帕累托控制机制,协助策略搜索多目标优先级最优解,为边缘云提供最优的资源调度策略,便于实时满足不同用户不同任务的资源需要,不仅使边缘云资源得到了充分利用,也大大提高了用户的使用体验。最后通过实验验证了该算法的有效性。

Mobile edge cloud computing is one of the cores of 5G technology,and it is also a very popular communication technology.However,the number of mobile users is growing rapidly,and the traditional resource allocation method can no longer meet user needs.Therefore,according to the real-time changes of user’s scale and task priority,how to reasonably formulate resource allocation strategies to meet user’s needs for computing units,storage space,software and other resources is a quite popular research direction at present.We propose an edge cloud resource scheduling algorithm based on multi-objective priority particle swarm algorithm,which rationally arranges multiple edge base stations to form an edge cloud.When multi-user and multi-task concurrently,integrating user data transmission rate,task energy consumption,task priority and edge base station performance and other factors,two fitness functions and a particle encoding and decoding method are designed,and Pareto control mechanism is introduced at the same time to assist the strategy search for the optimal solution of multi-objective priority and provide the optimal resource scheduling strategy for edge cloud,which is convenient to meet the resource needs of different users and tasks in real time.It not only makes full use of edge cloud resources,but also greatly improves user experience.Finally,the experimental results verify the effectiveness of the algorithm.