UNIVERSITA' DEGLI STUDI ROMA "TOR VERGATA"Intelligent software systems
The design and validation of complex, mission-critical systems require innovative methodologies that integrate Modeling & Simulation (M&S), Data-driven Engineering, and Artificial Intelligence within the broader framework of Model-based Systems Engineering (MBSE). In this context, the research activities of the Systems Engineering Lab (SEL) group are organized around the following areas.
Modeling & Simulation-Based Systems Engineering
M&S approaches are a key asset in the MBSE paradigm, particularly for the Verification & Validation (V&V) of complex and critical systems.
In this area, the group investigates advanced methodologies for the design and execution of local and distributed simulations, including the use of model transformation techniques and Artificial Intelligence to support and automate the different phases of the system lifecycle, from requirements analysis and design through to results verification in the context of simulation-based V&V, with a specific focus on extra-functional properties such as efficiency, availability, and reliability
Twinning Systems
A Digital Twin (DT) is a dynamic digital replica of a physical system or process that integrates simulation models and is continuously synchronized with its physical counterpart. The design and operation of systems that make use of digital twins (Twinning Systems) require approaches capable of effectively integrating heterogeneous, multi-scale architectures, formalisms, and technologies.
Research activities in this area include:
- the study of low-code systems engineering approaches for the automatic generation of DT models.
- the investigation of data-driven and AI-based methods for simulation results analysis, decision making, the definition of appropriate feedback mechanisms (closed-loop control) to support the adaptation, evolution, and optimization of the physical system;
Predictive Process Mining
Business Processes can be used to specify the behavior of complex systems ranging from distributed systems and manufacturing plants to entire organizations.
Such processes are often characterized by the pervasiveness of digital technologies (e.g., IoT nodes) and the generation and processing of large amounts of data. In this context, Predictive Process Mining is an innovative paradigm that combines Process Mining and M&S techniques.
The analysis of execution logs enables the derivation of data-driven simulation models, whose analysis in turn supports predictions on the future behavior of the process, enabling what-if evaluations, alternative scenario analysis, and the prediction and optimization of performance and reliability.
