Algorithm for 5G resource management optimization in edge computing

Abstract

The Internet of Things (IoT) brings new applications and challenges related to cloud computing. The service distribution challenge is becoming more evident and a need for better service options is emerging. The focus of the work is to optimize issues related to the allocation of resources in Edge Computing, improving the quality of service (QoS) with new methodologies. An algorithm based on a bio-inspired model was developed. This algorithm is based on the behavior of gray wolves and it is called Algorithm for 5G Resource Management Optmization in Edge Computing (GROMEC). The algorithm uses the meta-heuristic technique applied to Edge Computing, to result in a better allocation resources through user equipment (UE). The resources considered for allocation in that work are processing, memory, time and storage. Two genetic algorithms were used to define the fitness of an Edge in relation to the resource. Two other algorithms that use traditional techniques in the literature, the Best-First and AHP methods, were considered in the evaluation and comparison with the GROMEC. In the function used to calculate fitness during the simulation made with the GROMEC, the proposed algorithm had a lower number of denied services, presented a low number of blocks and was able to meet the largest number of UEs allocating on average up to 50% more in relation to the Best and 5.25% in relation to Nancy.