Resources allocation optimization algorithm based on the comprehensive utility in edge computing applications

Abstract

In the mobile edge computing environment, aiming at the problems of few classifications of resource nodes and low resource utilization in the process of multi-user and multi-server resource allocation, a resource optimization algorithm based on comprehensive utility is proposed. First, the algorithm improves the Naive Bayes algorithm, obtains the conditional probabilities of job types based on the established Naive Bayes formula and calculates the posterior probabilities of different job types under specific conditions. Second, the classification method of resource service nodes is designed. According to the resource utilization rate of the CPU and I/O, the resource service nodes are divided into CPU main resources and I/O main resources. Finally, the resource allocation based on comprehensive utility is considered. According to three factors, resource location, task priority and network transmission cost, the matching computing resource nodes are allocated to the job, and the optimal solution of matching job and resource nodes is obtained by the weighted bipartite graph method. The experimental results show that, compared with similar resource optimization algorithms, this method can effectively classify job types and resource service nodes, reduce resource occupancy rate and improve resource utilization rate.

Keywords

• edge computing
• resource allocation
• improved naive bayes algorithm
• resource service node classification
• weighted bipartite graph