Byzantine-Resilient Distributed Algorithm for Economic Dispatch: A Trust-Based Weight Allocation Mechanism

Abstract

This brief investigates the economic dispatch problem (EDP) in power systems under unreliable distributed networks. Potential Byzantine generator units (GUs) aim to disrupt the task by sending error messages. Normal GUs are dispatched to minimize total generation cost while adhering to generation constraints and maintaining supply-demand balance. To tackle this challenge, a resilient distributed algorithm is developed, incorporating locally computed-only auxiliary variables to accelerate convergence. Subsequently, a novel trust-based weight allocation mechanism (WAM) is introduced and integrated into the algorithm. Unlike traditional methods that rely on averaging neighbor information, this mechanism utilizes a well-designed negative trust function for aggregation. Theoretical analysis demonstrates that the proposed algorithm converges to the neighborhood of the optimal solution, and case studies on the IEEE 118-bus system illustrate the resilience of the algorithm against Byzantine attacks.