摘要：AbstractLarge-scale optimization in networked environment is challenged by both scalability and cyber-security. To address these challenges, this paper first develops an improved decentralized shrunken primal multi-dual subgradient (SPMDS) algorithm that is scalable with respect to both the agent fleet size and network dimension. This two-facade scalability is achieved by integrating the network topology reduction and multi-dual techniques into a single decentralized framework. Second, the homomorphic encryption is integrated into the SPDMS framework for privacy preserving. The novel cryptographic SPMDS adopts private key encryption to concurrently protect the privacy of agents and the system operator from adversaries. Efficacy of the proposed approaches will be validated against numerical examples of electric vehicle charging control over power distribution networks.
