Distributed optimization algorithms have many attractive features for coordinating systems with multiple agents. Using distributed optimization allows agents to collaborate in solving large-scale optimization problems while maintaining their autonomy. However, distributed algorithms may be vulnerable to cyberattacks due to their dependency on communication. Moreover, an actual distributed algorithm implementation requires a distributed termination method. This dissertation proposes a general cybersecurity-aware distributed optimization implementation framework for solving optimal power flow problems with a fault-tolerant distributed termination method.

This dissertation addresses the challenges of applying distributed optimization algorithms for solving optimal power flow problems in practical settings. We study several challenges, focusing on cybersecurity and the practical implementation of distributed optimization algorithms. This dissertation proposes a practical implementation framework to increase the resiliency of distributed optimization against cyberattacks.

This cybersecurity-aware distributed optimization implementation framework proposed in this dissertation includes (1) a framework to expedite the process of testing and experimenting with distributed optimization algorithms, (2) an empirical benchmark of multiple distributed algorithms with various optimal power flow models in the presence of nonideal communication, (3) cyberattack detection models via analyzing cyberattack threat models, (4) a mitigation strategy for cyberattacks and communication failures via formulating and solving a robust optimization problem, and (5) a fault-tolerant distributed termination method.