Geopolitical Collapse Simulator v3.0 — A Complex Adaptive System Digital Twin for Systemic Geopolitical Risk

The Geopolitical Collapse Simulator v3.0 is a reproducible complex adaptive system (CAS) digital twin designed to analyze how localized geopolitical shocks escalate into systemic failures across trade, energy, and information-cyber networks. The model integrates multi-layer network percolation, information entropy dynamics, heterogeneous strategic actors, and Multi-Agent Reinforcement Learning (MARL) to formalize collapse trajectories and post-collapse attractors. Monte Carlo experiments (N = 10,000) reveal a 0.732 systemic collapse probability and a 31.95% Distributed Autarky Transition rate. This v3.0 release includes the full research manuscript. Developed as an orchestration artifact of the O'Valley AI Orchestration Lab (OAOL).