区块链驱动的边缘计算系统联合优化算法

Joint optimization algorithm for blockchain driven edge computing systems

针对物联网设备在任务处理过程中能耗高及来自不同运营商的实体之间缺乏信任等问题,提出一种区块链驱动的边缘计算(Multi-access Edge Computing,MEC)系统联合优化算法。构建多接入边缘计算网络架构,其中基站作为边缘节点,同时兼具区块链子系统共识节点的功能。在可信性保证及用户服务质量保障的前提下,通过联合优化接入控制及计算资源分配策略,降低物联网设备的平均总能耗。考虑到网络的动态特性和信息获取的不确定性,将问题重建为马尔可夫决策过程,并采用基于深度确定性策略梯度(Deep Deterministic Policy Gradient,DDPG)的自适应决策算法求解。仿真结果表明,与随机接入算法和演员评论家算法相比,该算法能有效地降低网络能耗,提高DDPG智能体的长期平均累积奖励。

For the problems of high energy consumption during the task processing of Internet of Things(IoT)devices,and the lack of trust between entities from different operators,a joint optimization algorithm is proposed for a blockchain-driven multi-access edge computing(MEC)system.An MEC network architecture is built,in which base stations serve as edge nodes,while also serving the function of consensus nodes in the blockchain subsystem.With a focus on ensuring the trustworthiness and guaranteeing the quality of services,the algorithm aims to reduce the average total energy consumption of IoT devices through the joint optimization of access control and computing resource allocation strategies.Considering the dynamic nature of the network and uncertainties in information retrieval,the problem is reformulated as a Markov decision process,and an adaptive decision algorithm based on deep deterministic policy gradient(DDPG)is employed for problem solving.Simulation results indicate that,compared with the random access algorithms and the actor-critic algorithms,the proposed algorithm can effectively reduce network energy consumption,and can enhance the long-term average cumulative rewards of the DDPG intelligent agent.