Representation and Processing of Temporal Cases in Real-Time Intelligent Systems

The problem of improving the efficiency of decision making based on cases (case-based reasoning) in real-time intelligent systems is considered. Methods for preprocessing and storing temporal data are discussed. A method and algorithms for structuring the case base are proposed. The main stages of the method are the generalization of properties of dynamic parameters, the formation of classes of similar situations associated with cases from the case base, and the construction of decision trees to enable efficient search for solutions in each class. The results of a machine experiment are presented. To implement the proposed approach, a temporal database created using the Neo4j NoSQL graph database management system is used.