Reinforcement Learning for Computational Offloading in Fog-based IoT Systems: Applications, and Open Research Issues

Abstract

The fog computing has been widely integrated in the IoT-based systems allowing the fog nodes to offload and process tasks requested from IoT-enabled devices in a distributed manner instead of the centralized cloud servers to improve the systematic performance (i.e., reduced response delay, energy saving). However, designing the efficient offloading algorithms to achieve such the benefits is still challenging, mainly due to the complexity of fog computing environment and complicated requirements of computational tasks. Apart from many optimization, game theory, and heuristics based solutions, reinforcement learning (RL) is recently applied to provide intelligent offloading policies that can release the challenges efficiently. This paper presents an overview of RL applications to solve the computation offloading related problems in the fog computing environment. The open issues and challenges are explored and discussed for further study.