Cooperative Task Offloading and Block Mining in Blockchain-based Edge\n Computing with Multi-agent Deep Reinforcement Learning

The convergence of mobile edge computing (MEC) and blockchain is transforming\nthe current computing services in mobile networks, by offering task offloading\nsolutions with security enhancement empowered by blockchain mining.\nNevertheless, these important enabling technologies have been studied\nseparately in most existing works. This article proposes a novel cooperative\ntask offloading and block mining (TOBM) scheme for a blockchain-based MEC\nsystem where each edge device not only handles data tasks but also deals with\nblock mining for improving the system utility. To address the latency issues\ncaused by the blockchain operation in MEC, we develop a new Proof-of-Reputation\nconsensus mechanism based on a lightweight block verification strategy. A\nmulti-objective function is then formulated to maximize the system utility of\nthe blockchain-based MEC system, by jointly optimizing offloading decision,\nchannel selection, transmit power allocation, and computational resource\nallocation. We propose a novel distributed deep reinforcement learning-based\napproach by using a multi-agent deep deterministic policy gradient algorithm.\nWe then develop a game-theoretic solution to model the offloading and mining\ncompetition among edge devices as a potential game, and prove the existence of\na pure Nash equilibrium. Simulation results demonstrate the significant system\nutility improvements of our proposed scheme over baseline approaches.\n