Development of knowledge representation based on markov logical networks in the business process mangement system

Abstract

The problem of constructing knowledge representation in the process control system based on the analysis of the behavior of business processes, represented in the form of logs of events, is studied. Each event characterizes the action of the business process. The urgency of the problem is determined by the fact that when managing complex knowledge-capacious business processes, performers can change the sequence of actions taking into account additional knowledge about the subject area. As a result, there is a discrepancy between the process and its model, which creates difficulties for the further management of this business process. To eliminate this discrepancy, it is necessary to formalize the additional knowledge used and apply them in process management, which requires the creation of an appropriate knowledge representation. The proposed knowledge representation model takes into account the static and dynamic characteristics of the business process. The static characteristics of a business process are specified by facts and rules with arguments represented by the attributes of the log events. Facts and rules are formed on the basis of appropriate templates. Attributes specify the values of the properties of objects with which the business process operates. Dynamic features of the business process are determined through the current distribution of the probability that the rules will be executed, taking into account the attributes of the current business process log event. The proposed model is characterized by the fact that it takes into account the limitations on the permissible sequences of execution of the actions of the business process, as well as restrictions based on a priori knowledge of the subject area. Such restrictions will reduce the complexity of the problem of finding the probabilities of a successful completion of a business process by reducing the number of allowed trails in the event that the performers have changed the sequence of actions. In practical terms, the model provides the ability to support decision-making on the management of knowledge-intensive business processes based on predicting the probabilities of achieving the final state of the process, taking into account the attributes of log events.