Relationship between resource scheduling and distributed learning in IoT edge computing — An insight into complementary aspects, existing research and future directions

Abstract

Resource Scheduling and Distributed learning play a key role in Internet of Things (IoT) edge computing systems. There has been extensive research in each area, however, there is limited work focusing on the relationship between the two. We present a systematic literature review (SLR) examining the relationship between the two by thoroughly reviewing the available articles in these two specific areas. Our main novel contribution is to discover a complementary relationship between resource scheduling and distributed learning. We find that the resource scheduling techniques are utilized for distributed machine learning (DML) in edge networks, while distributed reinforcement learning (RL) is used as an optimization technique for resource scheduling in edge networks. Other key contributions of the SLR include: (1) presenting a detailed taxonomy on resource scheduling and distributed learning in edge computing, (2) reviewing articles on resource scheduling for DML and distributed RL for resource scheduling, mapping them to the taxonomy, and classifying them into broad categories, and (3) discussing the future research directions as well as the challenges arising from the integration of new technologies with resource scheduling and distributed learning in edge networks.