Business processes monitoring based on fuzzy cognitive maps

Abstract

The subject matter of the article is a toolkit and monitoring processes of weakly structured business processes using fuzzy cognitive maps (FCMs). The goal is to create a model for monitoring and forecasting the course of business processes based on the FCMs, which provide flexibility and the ability to adapt to the conditions of changing circumstances in the process to be monitored, as well as insurance of the possibility of applying unstable FCM.s The task is to develop a formal monitoring system model; develop a computational FCM model; develop a method of using unstable FCMs: to develop a method for creating and using an FCM model. The methods used are graph modeling methods and computational experiment methods. The following results were obtained. A formal model of the system of monitoring and coordination of decisions regarding the course of business processes has been developed. A model of a weighted semantic graph was developed, which includes three types of vertices, namely, input vertices that will receive subjective evaluations of users regarding the course of the process, intermediate vertices that correspond to factors important from the perspective of the decision-maker, and final vertices that reflect integral evaluations of the quality of business process execution. A computational FCM model has been developed, which can adapt to the specifics of business logic due to the flexible adjustment of memory parameters and connections between FCM nodes. A method of using unstable FCMs has been developed. A methodology for creating and using FCMs designed for monitoring weakly structured business processes has been developed. Conclusions. The scientific novelty of the work is as follows: the model for monitoring and forecasting the course of business processes due to the use of a fuzzy cognitive map, in which the parameters of the importance and node memory  and connections provide additional flexibility and the possibility of coordination and adaptation in the conditions of changing circumstances in the process to be monitored has been improved; a method of using unstable FCMs by setting limits on the values of the nodes excitation and using as a measure of excitation not only stable excitation values of FCM nodes but also the rate of growth of excitation values as an indicator of trends, which allows monitoring and forecasting the course of business processes has been proposed.