Research on Dynamic Deduction of Plans Based on Knowledge Graph
Zhiyong Li, Huiming Guo, Mengqiao Zhang · 2022 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA) · 2022
As the guiding text of emergency response, the plan is unstructured information, and its existing static management mode cannot dynamically adapt to the deduction needs of different scenarios. To this end, this paper proposes a dynamic derivation method based on knowledge graph: for large-scale event emergencies, sorting out knowledge of business scenarios, designing and constructing domain knowledge graphs including geographic information ontology, contingency text ontology, and deduction scene ontology. Moreover, use Neo4j to store and manage the constructed knowledge graph. Based on this, the typical scenario of “crowded and stampede event in security check area” is selected. Based on the FSR (Force-Scene-Response) model theory, the inference rules are constructed, the deduction model is designed, and the model is simulated and analyzed. The simulation results show that The disposal instructions generated by the model can effectively mitigate risk situations. Finally, a dynamic deduction auxiliary decision-making system with good interactivity is developed. Taking the crowded stampede event in the security check area as an example, the system generates the disposal strategy and the situation evolution process according to the input parameters, verifying the validity of the proposed model and method. It shows that the system can adapt to different scenarios and support dynamic deduction of plans.