Modelagem e simulação de sistemas de gerenciamento de processos de negócios baseadas em workflow net temporais com mecanismos de alocação de recursos híbridos fuzzy

Abstract

The main objective of this doctoral research is to propose a new business process model for efficient (and realistic) resources and time management in Workflow nets. This research had as its starting point the validation of the fuzzy resource allocation model presented in (JESKE, 2006). In particular, a specialized inference mechanism of the token player type adapted to a fuzzy hybrid resource allocation model was implemented in the modeling, analysis and simulation tool of colored Petri nets called CPN Tools. Such simulations have shown that the new model in relation to a simple resource allocation mechanism with an earlier firing policy (FIFO) allows an improvement in the proportion of cases that meet the established deadlines. To solve the scheduling problem in Workflow Management Systems, important characteristics inherent to these systems should be considered. In particular, when many cases are run concurrently, conflict situations due to shared resource can occur and must be resolved in real time (without a backward mechanism). Also, temporary restrictions regarding the deadlines for delivery of specific cases should be considered. For that, a fuzzy time associated to the activities of the processes was defined as well as a mechanism of propagation of restrictions whose role is to guarantee the consistency of the time constraints to carry out the activities respecting the expected delivery times. With the objective of making the behavior of the new model the closest and consistent with the behavior of the human collaborators involved in the execution of business process activities, the notion of joint possibility was inserted in the model. This notion involves the uncertain and nebulous character of the execution times of the activities of the processes and the intensity of the resources used to carry out the activities. New firing rules for transitions were then defined based on the concept presented. Token players based on the new model presented were defined, implemented, simulated and validated in the CPN Tools, whose main role is to show the execution of the new models presented. Using the CPN Tools, it was possible to evaluate the system behavior considering the expected or most critical operating. It is important to point out that through the simulation results it is possible to predict how the system will behave and, from this, to evaluate several possible resource distribution scenarios in order to be able, in the future, to carry out the sizing of the system based on a modeling and simulation tool that approximates of the way employees work in an office environment.