Optimizing the Resilience of the 3E System: A Coupled Supernetwork–ABM Framework for the Yangtze River Delta

The transition toward carbon neutrality demands not only an understanding of the complex dynamics within energy–economy–environment (3E) systems but also the ability to strategically enhance their resilience against external shocks. Here, we introduce a bidirectional coupled framework that integrates macro-level supernetwork topology with micro-level agent-based modeling (ABM) to diagnose structural vulnerabilities and optimize systemic performance. Applied to 41 cities in the Yangtze River Delta (YRD) from 2010 to 2023, our framework reveals a persistent core–periphery spatial disparity in coupling coordination, underpinned by four distinct network layers—energy flow, economic linkage, environmental impact, and policy synergy—whose densities vary by an order of magnitude. Through coupled evolutionary simulations, we quantify system resilience as a tripartite metric of robustness, adaptability, and recovery, identifying a systemic structural weakness: a robust recovery capacity is offset by substantially lower robustness. To address this, we deploy a genetic algorithm to solve for optimal investment allocation under a budget constraint, demonstrating that a targeted, hub-centric strategy yields a higher marginal resilience gain than uniform distribution. Furthermore, embedding multi-agent reinforcement learning (MARL) and social learning into policy scenario simulations shows that unified environmental standards and a carbon-inclusive mechanism effectively eliminate regulatory arbitrage and accelerate low-carbon behavioral diffusion, improving system-wide robustness by 35% under extreme climate shocks. This work delivers a closed-loop, transferable analytical framework that transforms structural diagnosis into actionable optimization, offering a scientific basis for coordinated regional decarbonization strategies.