Electric Vehicle Charging Networks (EVCNs) are increasingly vulnerable to disruptions from the power grid. Current approaches for post-outage Quality of Service (QoS) recovery rely on fixed partial charging strategies or fuzzy logic inference systems, which lack adaptability to complex dynamic conditions. This paper introduces a novel hybrid modeling approach that integrates System Dynamics (SD) with Agent-Based Modeling (ABM) to simulate and optimize EVCN QoS recovery following prolonged outages.

The proposed framework leverages SD modeling to capture critical interdependencies between power grid components, charging demand, and network performance metrics. These interdependencies form feedback loops that significantly influence recovery trajectories but are often overlooked in traditional approaches. Concurrently, individual EV charging behaviors and station operations are modeled using ABM, representing heterogeneous decision-making processes of EV owners and station managers. Geographic Information System (GIS) visualization enables spatial analysis of congestion patterns and QoS recovery strategies across urban districts.

We evaluate our hybrid model using the modified Roy Billinton test system with six interconnected districts experiencing a 31-hour outage scenario. Results demonstrate that the SD-ABM approach achieves 23% faster recovery in grid voltage stability and 34% reduction in charging station congestion compared to fixed mitigation strategies and fuzzy logic inference systems. The model reveals emergent spatial patterns in EV redistribution and identifies previously unrecognized interdependencies between charging station utilization, grid stability, and user experience metrics.

Key contributions include: (1) a methodological framework for integrating system-level dynamics with agent-level behaviors in EVCN QoS recovery planning; (2) identification of critical interdependencies and threshold effects that trigger cascading congestion; and (3) spatiotemporal optimization strategies for maintaining acceptable QoS during recovery periods. This approach offers practical decision support for grid operators and urban planners seeking to enhance EVCN resilience with adaptive and computationally efficient modeling techniques.