Privacy-Preserving Decentralized Multi-Agent Cooperative Optimization -- Paradigm Design and Privacy Analysis

TL;DR

This paper introduces a novel privacy-preserving decentralized optimization framework for multi-agent systems using Paillier cryptosystem, ensuring security and correctness while resisting attacks, validated through simulations and real-world tests.

Contribution

It presents a new decentralized optimization paradigm that guarantees privacy and security using cryptography, addressing gaps in existing methods.

Findings

Achieves ideal correctness and security in decentralized optimization.

Resists a range of adversarial attacks.

Validated through numerical simulations and real-world experiments.

Abstract

Large-scale multi-agent cooperative control problems have materially enjoyed the scalability, adaptivity, and flexibility of decentralized optimization. However, due to the mandatory iterative communications between the agents and the system operator, the decentralized architecture is vulnerable to malicious attacks and privacy breach. Current research on addressing privacy preservation of both agents and the system operator in cooperative decentralized optimization with strongly coupled objective functions and constraints is still primitive. To fill in the gaps, this paper proposes a novel privacy-preserving decentralized optimization paradigm based on Paillier cryptosystem. The proposed paradigm achieves ideal correctness and security, as well as resists attacks from a range of adversaries. The efficacy and efficiency of the proposed approach are verified via numerical simulations and a real-world physical platform.