A Computational Offloading Method Based on Resource Joint Optimization

With the advent of 5G era and the continuous development of mobile Internet, various mobile terminal tasks are gradually diversified and intensive. Ordinary mobile devices can not meet the large-scale data processing scenarios and need more computing and storage resources to deal with complex tasks. Cloud data centre has strong computing resources and storage capacity. The mobile edge network can be formed by migrating the computing resources of cloud data centre to the network edge side. By comprehensively considering the parameters of task type, task size, communication environment, server resources and server power, the resources are jointly allocated and the reasonable offloading decision is made. The computing tasks are offloaded to the edge server, so as to achieve the effect of saving energy, reducing delay and improving efficiency. This paper mainly uses clustering algorithm and KM algorithm to develop the offloading algorithm. Firstly, the terminal tasks are classified by clustering algorithm and the appropriate terminal tasks are selected to match with the edge server. Secondly, the KM algorithm is used for bipartite graph matching to find the appropriate edge server for each terminal task to offload, then the overall offloading cost is minimized. Finally, the experimental results verify the effectiveness of the proposed method in terms of offloading cost and iteration cost.