In the scenario of a large-scale power outage after an extreme disaster

such as a severe ice storm

the distribution system with multiple distributed generations (DGs) is of great value for post-disaster load restoration. However

due to the uncertainty of renewable energy output and the controllability of different DGs

effective utilization of these DGs becomes an urgent issue. To address the uncertainty of renewable energy output under disasters

this paper proposes a multi-stage optimization restoration strategy for a distribution system with distributed resources

such as a mobile energy storage system (MESS)

integrated energy system (IES)

and photovoltaic (PV). In particular

this study extracts historical data features by utilizing improved conditional generative adversarial nets (CGAN) to generate PV output scenarios. Subsequently

according to the dynamic and static characteristics of the power supply

the time sequence model of each distributed resource is established. On the premise of meeting the constraints of emergency microgrids and load characteristics

the optimal MESS configuration scheme and controllable DGs output are achieved to maximize the restoration of the power supply for critical loads. Finally

the case studies of IEEE 33-bus and PE&G 69-bus systems demonstrate the effectiveness of the proposed model in enhancing the resilience of the distribution system.