-To model business processes, a variety of modeling tools and techniques are used. Colored Petri Nets (CPN) are one of the modeling notations capable of presenting different patterns in activity flows, e.g., concurrency of activities in the process. CPN Tools enables the creation and execution of business process simulation models based on CPN notation and Standard Machine Language (SML) w.r.t. different aspects of processes. For instance, assigning different resources to different tasks, generating attributes for cases, and modeling stochastic Petri nets for choices are possible. We can replay and capture process behavior in the form of event logs for what-if analyses using simulation models of business processes. However, modeling the processes and implementing SML functions for generating event logs are complicated. We designed and implemented a tool that extracts process parameters directly from an event log and converts the event log using process mining techniques to CPN models along with generated SML functions. The tool allows for the straightforward modeling of business processes without user interaction, as well as the ability to change via a user interface.
Other techniques propose the generation of process simulation
models without the use of CPN Tools and in the form of
independent platforms, such as [
In this paper, we introduce a new tool that is able to generate a ready to simulate CPN models directly from an event log. Our tool (GenCPN) is a web application based on Python, which reduces all the user interaction and efforts for designing the models from scratch. Moreover, the process mining insights are put into action to extract all the required process simulation parameters. The strong contribution of this tool is providing the implemented SML functions that directly include all the interactions of the simulation engine (daemon) for the CPN models. It also provides the ready-to-run CPN models in CPN Tools. with the extracted arrival rate from the event log. For logging generated events in the form of an event log, the function create log file in the Openfile transition is designed (Figure 3 (a)). Figure 3 (b) represents the generated function for executing each activity based on their extracted duration from the event log, as well as a sample scenario of sharing resources (resource pool) between two activities.
CPN Tools is being used for both academic and industrial
purposes. GenCPN which generates CPN models
automatically based on the event logs of processes is used for
generating event logs of different projects such as IOP project1
to regenerate event logs of production lines in a different
setting. Moreover, different simulation models are designed
in the previous and current researches, e.g., [
GenCPN is publicly available in the form of web application2 along with all the codes and the screencast as open source3 for further uses. Figure 4 is part of the tool’s user interface, where the enriched discovered CPN model from the event log is shown (left). The discovered parameters are also shown (right) and possible to be changed by the user if required, e.g., arrival rate or an activity service time. The output of the tool is a zip file including the CPN file and the corresponding SML file, which is ready to be stimulated via the graphical interface of the CPN Tools.
Designing executable and close-to-reality prescription models is a challenging task. Process mining is used to generate and design these prescribing models as close to reality as possible. We presented a tool in this paper that automatically converts event logs into executable simulation models. The simulation model parameters are filled in with real values from previous process executions. The model is created as an enriched CPN model, which can be run using CPN Tools, a sophisticated tool for simulating Petri nets. The required SML functions are also generated together with the model in order to capture simulation results in the form of event logs and to allow the user to make additional changes while running the simulation, e.g., adjusting the working hours or the density of arrival rate for different months.
Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 2023 Internet of ProductionProject ID: 390621612. We also thank the Alexander von Humboldt (AvH) Stiftung for supporting our research.