Simplificação de modelos de processos utilizando propriedades de autômatos e modularização

Abstract

In the context of computer science, we can define a process as a set of activities that takes place in a logic sequence to obtain results. The business process management (BPM) is the set of tecnics, tool and methods used to support the phases of life cycle of a business process, bringing several benefits to its user, such as simulation, documentation and a greater understanding of the processes. Process mining uses event logs, i.e, sets of activities, to extract knowledge, being able to be represented in many ways, such as business process management notation (BPMN), Petri nets or transitions systems, such as automata, for example. When we represent complex event logs, with a lot of process, through automata, it is verified that there are state explosion problem, generating an intelligible model. This work proposes an approach that uses automata decomposition concepts to modularise processes. In the preliminary experiments, we present promising results in the modularisation of a simple automata, but in the future, minimal automata generated from a real event log provided by the Ceará State Finance Department (SEFAZ-CE) will be used.