Abstract:

In order to improve the coordination performance of multi-agents in emergency rescue, the network relationship and internal correlation of four agents were sorted out for urban flood disaster events. Combined with the multi-agent theory, the multi-agent interaction and cooperation framework was built. The multi-agent cooperative simulation model for urban flood disaster emergency rescue was designed. The changes in the number of rescued victims, average rescue time and total rescue time under different disaster scenarios were studied. Results show that the extent of the disaster, the scope of search, the ability of the dominant rescuer and the number of main rescuer are important factors influencing the level of rescue collaboration. Enhancing the scope of search, improving the ability of the dominant rescuer and increasing the investment of the main rescuer contribute to the improvement of the synergistic capacity of disaster emergency response, and the proportion of those affected by disasters who are rescued rises. The robustness analysis shows that the scope of cooperation has a small effect on the number of rescued victims, and the scope of search has a significant effect on the number of rescued victims and the total rescue time. The sensitivity analysis shows that the ability of the dominant rescuer and the number of main rescuer both have a critical value. Once these the critical values are exceeded, the rescue effect will tend to be stabilized.

Key words: urban flood disaster, emergency rescue, multi-agent cooperation, multi-agent theory, rescue efficiency