Benchmark Datasets for Adversarial Distributed Optimization #1

This repository contains the benchmark datasets used in the study "Optimization under Attack: Resilience, Vulnerability, and the Path to Collapse". The datasets support the evaluation of adversarial distributed optimization in discrete-choice multi-agent systems across three application domains: energy demand, privacy-preserving data sharing, and synthetic Gaussian scenarios. The benchmark includes datasets generated from extensive large-scale experiments that systematically vary adversarial scale, behavioural severity, and network position to analyse system resilience, vulnerability, collapse, and Pareto optimality. The repository contains the Energy, Privacy (high- and low-signal), and Gaussian datasets. The Voting dataset is available in a companion Zenodo repository, and both records are cross-linked to provide access to the complete benchmark collection. These datasets are intended to support reproducible research, benchmarking, and future studies on adversarial distributed optimization, multi-agent systems, resilience analysis, and fault-tolerant optimization.