Simulator of Information System Behavior based on Two Notions of Object-Centric Petri Nets

The problem of object-centric nature of business processes has been recognized by researchers in process mining community with standardization of object-centric event-logs format (OCEL) as well as definition of Object-Centric Petri nets (OC-net) and implementation of object-centric discovery algorithms. However, there are still no software tools that enable both business process modeling in OC-net notation and the generation of object-centric event-logs, which can be used to improve algorithms through the use of artificial event-logs as input. OC-net notation offers tangible benefits for modeling and simulation - it provides high-level view of a process with relationships between objects of different kinds, whilst requiring no explicit synchronization rules for various tokens passing through same business process activity. The active development of object-centric process mining algorithms calls for tools to improve and test these algorithms. Typically, such tools generate artificial event-logs with adjustable parameters that researchers can modify. This project aims to build an OC-net simulation software to model and simulate business processes, and generate object-centric event logs that can be used to enhance existing object-centric process mining algorithms. One of the stakeholders of the proposed project is Laboratory of Process-Aware Information Systems at the HSE University’s Faculty of Computer Science (PAIS Lab), where this work is initiated. The system is shipped as a stand-alone application with a decoupled architecture for ease of reuse, and developed with a defined scope, requirements, and evaluation plan. The outcome is a software system that meets the needs of both research and model development communities. The latest release with documentation can be accessed at https://github.com/MisterPotz/ocgena. Key Words: Object-Centric Petri net; business process modelling; simulation; object-centric event-log; event-log generation