Resilient Distributed Optimization Algorithms for Resource Allocation
Resilient Distributed Optimization Algorithms for Resource Allocation
复制标题
DOI：
10.1109/cdc40024.2019.9030051
复制
发表时间：
2019-04
期刊：
影响因子：
--
通讯作者：
César A. Uribe;Hoi-To Wai;M. Alizadeh
中科院分区：
文献类型：
--
作者：
César A. Uribe;Hoi-To Wai;M. Alizadeh
分布式算法为资源分配问题（例如，网络物理系统）提供了比集中式算法更大的灵活性。然而，这些算法的分布式特性经常使系统容易受到中间人攻击，特别是当消息在定价代理和中央协调器之间传输时。在原始-对偶分布式优化框架下，提出了一种分布式算法的弹性策略。我们制定了一个健壮的优化模型，该模型考虑了对代理和协调器之间通信通道的拜占庭攻击。我们提出了一个弹性的原始对偶算法使用最先进的鲁棒统计方法。结果表明，该算法收敛于鲁棒优化模型的一个邻域，该邻域的半径与受攻击信道的比例成正比。
Distributed algorithms provide flexibility over centralized algorithms for resource allocation problems, e.g., cyber-physical systems. However, the distributed nature of these algorithms often makes the systems susceptible to man-in-the-middle attacks, especially when messages are transmitted between price-taking agents and a central coordinator. We propose a resilient strategy for distributed algorithms under the framework of primal-dual distributed optimization. We formulate a robust optimization model that accounts for Byzantine attacks on the communication channels between agents and coordinator. We propose a resilient primal-dual algorithm using state-of-the-art robust statistics methods. The proposed algorithm is shown to converge to a neighborhood of the robust optimization model, where the neighborhood’s radius is proportional to the fraction of attacked channels.