MODEL OF GENERALIZED REPRESENTATION OF TEMPORAL KNOWLEDGE FOR TASKS OF SUPPORT OF ADMINISTRATIVE DECISIONS

Abstract

The subject of the research is the processes of building and using temporal knowledge, which determine the order of implementation of managerial decisions within the framework of organizational management. The goal is to develop a temporal knowledge representation model to determine the sequence of control actions over time as part of a management decision, taking into account the level of the organizational hierarchy at which these actions are implemented. To achieve this goal, the problems of structuring the factors determining the differences in the presentation and use of temporal knowledge, as well as the development of a generalized model of temporal knowledge to support managerial decisions, are solved. It is shown that the key differences between temporal and causal dependencies are associated with the use of a combination of formal and informal knowledge of performers about causal relationships between control actions as part of a management decision to build a sequence of actions in time, as well as the probabilistic properties of temporal relationships. A generalized model of temporal knowledge is proposed, designed to support managerial decisions. The model includes many generalized facts, as well as a set of temporal and hierarchical relationships between these facts. Facts reflect known states of the control object. Each generalized fact is formed on the basis of a conjunction of elementary facts. An elementary fact sets the value of a certain property of a control object. Temporal relationships determine the sequence of facts over time. The organizational hierarchy level for facts is set using logical operations. In practical terms, the proposed model is focused on building a temporal knowledge base. Conclusion based on such a knowledge base makes it possible to determine the probable sequence of control actions that lead to a sequence of transitions between states, which allows achieving the target state of the control object. This makes it possible to form a set of possible alternative options for implementing a managerial decision in the event of abnormal conditions of the control object.