Scientific publications

The ASIMOV project investigates the combination of Digital Twins (DTs) and Artificial Intelligence (AI) to find the opportunities and challenges for automated optimization and calibration of complex high-tech systems in complex environments. In many cases the actual system is not available for training AI components, therefore a dedicated digital twin or digital model is set up for providing that training data.

This document describes fundamental considerations towards a process for identification of relevant parameters for a digital twin. The process includes identifying parameters in the real-world product (physical twin) which are relevant for its digital twin given a specific use case and describing methods to acquire and use these parameters.

The optimization of complex cyber-physical systems is a crucial task for their correct functioning, usability, and commercial viability. Due to their complexity, scale and resource intensiveness, conventionalmanual optimization is infeasible in many instances. We investigate the combination of the Digital Twin paradigm and Reinforcement Learning framework to address the long response times, limited availability ofdata, and the intractability of such systems.

Compositionality supports the manipulation of large systems by working on their components. For model-based testing, this means that large systems can be tested by modelling and testing their components: passing tests for all components implies passing tests for the whole system. In previous work [14], we defined mutual acceptance for specification models and proved that this is a sufficient condition for compositionality in model-based testing.

We combine Behavior-Driven Development (BDD) and formal Model-Based Testing (MBT), to benefit from the smooth collaboration among stakeholders in BDD, and from automated testing with precise test cases in MBT. However, textual BDD scenarios written in natural language are not sufficient for formal MBT, as they might be ambiguous and lack the required information for testing.

In today’s digital society, diverse computing de-vices-from edge devices to data centers-support various applications, each with specific performance and energy characteristics. Analyzing application energy consumption is crucial for improving energy efficiency, optimizing resources, and reducing environmental impact.

This is the artifact accompanying the paper: Dennis Hendriks and Wytse Oortwijn, "gLTSdiff: A Generalized Framework for Structural Comparison of Software Behavior", submitted to the International Journal on Software and Systems Modeling (SoSyM) in 2024. This artifact includes all models, code and scripts needed to reproduce the results from the paper.

Large, complex systems are often divided into components. As these systems and their software inevitably evolve, engineers must understand the (communication) behavior of the software to properly change it, and understand the impact of their changes to prevent costly regressions and reduce risks. Creating an overview of the software behavior is challenging, time-consuming and error-prone, even with the available code, tests and documentation.

Supervisor synthesis is a means to algorithmically derive a supervisory controller from a discrete-event model of a system and a requirements specification. For large systems, synthesis suffers from state space explosion. To mitigate this, synthesis can be applied to a symbolic representation of the models by using Binary Decision Diagrams (BDDs).

For software maintenance and evolution, insight into the codebase is crucial. One way to enhance insight is the application of static analysis to extract and visualize program-specific relations from the code itself, such as call graphs and inheritance trees. Yet, software often contains in-house libraries: unique, domain-specific libraries whose usage is typically scattered throughout the codebase.