Generation of agent simulation models by using process mining methods on the example of E-learning process

Abstract

The main objective of this work is to develop a method for the formation of agent simulation models based on the actual business processes implementation data. To solve this problem, we used the educational process management system Moodle, ProM process mining system, and AnyLogic simulation system. The initial data of the educational process obtained from the Moodle system were processed using Process Mining methods, and a heuristic miner based on the Weijters algorithm was used. Business process models were described in the eEPC notation. The prognostic ability of simulation models constructed as a result of the application of the proposed method turned out to be quite high. The error in predicting the total execution time of the process and the total complexity of the process did not exceed 7%.