Supplementary information for Fast and secure computational offloading with Lagrange coded mobile edge computing

Article abstractThis paper proposes a novel framework based on Lagrange coded computing (LCC) for fast and secure offloading of computing tasks in the mobile edge computing (MEC) network. The network is formed by multiple base stations (BSs) acting as 'masters' which offload their computations to edge devices acting as 'workers'. The framework aims to ensure efficient allocation of computing loads and bandwidths to workers, and providing them with proper incentives to finish their tasks by the specified deadlines. Thus, each master must decide on the amounts of allocated load and bandwidth, and a service fee paid to each worker given that: i) other masters, i.e., BSs, can be privately-owned or controlled by different operators, i.e., they do not communicate/coordinate their decisions with the master; ii) workers are heterogeneous non-dedicated edge devices with constrained and nondeterministic computing resources. As such, masters compete for the best workers in a stochastic and partially-observable environment. To describe interactions between masters and workers, we formulate a new stochastic auction model with contingent values of bidders, i.e., masters and contingent payments to auctioneers, i.e., workers. To solve the auction, we represent it as a stochastic Bayesian game and develop machine learning algorithms to improve the auction solution.© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

论文摘要
本文提出一种基于拉格朗日编码计算（Lagrange coded computing, LCC）的新型框架，用于移动边缘计算（mobile edge computing, MEC）网络中计算任务的快速安全卸载。该网络由多个充当‘主节点’的基站（base stations, BSs）构成，这些主节点将其计算任务卸载至充当‘工作节点’的边缘设备。该框架旨在确保向工作节点高效分配计算负载与带宽，并为其提供适当激励，促使其在指定截止期限前完成任务。因此，每个主节点需决定分配给各工作节点的负载量、带宽量及服务费用，约束条件包括：i) 其他主节点（即BSs）可能为私有或由不同运营商控制，即它们不会与该主节点进行决策沟通或协调；ii) 工作节点为异构非专用边缘设备，其计算资源受限且具有不确定性。因此，主节点需在随机且部分可观测的环境中竞争最优工作节点。为描述主节点与工作节点间的交互，我们构建了一种新的随机拍卖模型，其中投标人（即主节点）具有或有价值，且需向拍卖人（即工作节点）支付或有款项。为求解该拍卖模型，我们将其表示为随机贝叶斯博弈，并开发机器学习算法以优化拍卖方案。
© 2021 IEEE。允许个人使用本材料。所有其他用途（包括在任何当前或未来媒体中出于广告或推广目的重印/再版本材料、创建新的集体作品、转售或重新分发至服务器或列表，或在其他作品中重用本作品的任何受版权保护的组件）均需获得IEEE的许可。