Verification of Digital Twins using Classical & Statistical Model Checking
Raghavendran Gunasekaran, Boudewijn Haverkort · 2025
Read Verification of Digital Twins using Classical & Statistical Model Checking by Raghavendran Gunasekaran, Boudewijn Haverkort (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-04322.txt.
Introduction
Digital twins shadow real machines with a living bundle of models, and Raghavendran Gunasekaran and Boudewijn Haverkort ask how to keep the shadow honest. Because the virtual copy updates constantly and its parts interact in random, time-critical ways, it can drift into inconsistent or unsafe behavior at runtime. Checking it once at design time is not enough.
The authors pair two checking styles. Classical model checking proves hard properties like freedom from deadlock, while statistical checking samples the random behavior twins actually show. They report lessons from applying both to the twin of a self-driving truck. The combined approach catches both logical flaws and timing or probability problems that either method alone would miss.
Catalog record
| CREATOR | Raghavendran Gunasekaran, Boudewijn Haverkort |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-04322
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Verification of Digital Twins using Classical & Statistical Model Checking Raghavendran Gunasekaran, Boudewijn Haverkort Abstract With the increasing adoption of digital techniques, the concept of digital twin (DT) has received a widespread attention in both industry and academia. While several definitions exist for a DT, most definitions focus on the existence of a virtual entity (VE) of a real-world object or process, often comprising interconnected models which interact with each other, undergoing changes continuously owing to the synchronization with the real-world object. These interactions might lead to inconsistencies at execution time, due to their highly stochastic and/or time-critical nature, which may lead to undesirable behavior. In addition, the continuously varying nature of VE owing to its synchronization with the real-world object further contributes to the complexity arising from these interactions and corresponding model execution times, which could possibly affect its overall functioning at runtime. This creates a need to perform (continuous) verification of the VE, to ensure that it behaves consistently at runtime by adhering to desired properties such as deadlock freeness, functional correctness, liveness and timeliness. Some critical properties such as deadlock freeness can only be verified using classical model checking; on the other hand, statistical model checking provides the possibility to model actual stochastic temporal behavior. We therefore propose to use both these techniques to verify the correctness and the fulfillment of desirable properties of VE. We present our observations and findings from applying these techniques on the DT of an autonomously driving truck. Results from these verification techniques suggest that this DT adheres to properties of deadlock freeness and functional correctness, but not adhering to timeliness properties. 1 Introduction With the advent of cloud computing, big data and artificial intelligence, the process of digitalisation takes the center stage for driving innovation. One of the key technologies for digitalisation gaining widespread attention in both industry and academia are digital twins (DTs). They have been used for several applications in industry, e.g., control, monitoring, predictive maintenance, validation, design. Since its inception in 2002, DTs have been provided with several definitions [ 24 , 14 , 13 , 9 , 26 , 17 ] ; most of these definitions focus on the existence of a virtual replica of a physical object or process (often called actual entity (AE)), which comprises several interconnected models and which synchronizes continuously with its physical counterpart. These models interact with each other to exchange data in order to effect the purpose of the DT during its execution. The myriad of interactions within a DT could cause inconsistent behavior at runtime which is undesirable [ 27 ] . In addition, the continuous synchronization between the VE and AE contributes to a continuous evolution of the DT during its lifecycle, which poses additional complexity at runtime; requiring continuous verification of DTs. To the best of our knowledge, there exists no literature which explicitly addresses the verification and validation (V&V) of runtime behavior of DTs. In this paper, we propose to combine the techniques of classical model checking (CMC) and statistical model checking (SMC) to verify the run-time behavior of DTs. We apply these techniques to a DT of an autonomously driving truck (which lacks complete or structured documentation) to verify the desired properties of interest using the UPPAAL toolset [ 7 ] . The sections on the DT case study and behavioral modeling of this DT in UPPAAL has already been covered in our previous work [ 16 ] . Due to space constraints, we omit these sections in this paper and we refer the readers to our previous work, Gunasekaran and Haverkort [ 16 ] . Moreover, from our observations and findings, we compare and contrast SMC with CMC to guide the reader in selecting the right technique for verification of their DT. 2 Background The runtime behavior of a DT is the emergent behavior arising from the compositional effect of interactions between its different components. Tao et al. [ 25 ] and Van den Brand et al. [ 27 ] point to the runtime complexity of the multitude of interactions (which could be stochastic and time-critical in nature) occurring within a DT which could lead to inconsistent or under-performing behavior, both of which are undesirable. This type of behavior can only be observed during the execution of the overall DT, i.e., when the different components in the DT interact with each other; it cannot be determined from the individual components. In order to completely avoid such undesirable behavior, verification of the runtime behavior of the DT can come to the rescue. In addition to the complexity arising from multitude of interactions between cross-domain models, the evolution of VE as elucidated by Zhang et al. [ 28 ] and Mertens et al. [ 19 ] due to its continuous synchronization with AE and other reasons pose further challenges. This variability in VE across its lifecycle, results in the adaptation of its models and the interactions between the models, leading to further uncertainty issues at runtime. Thus, performing verification of VE only at its design stage does not assure its correct functioning. Consequently, the dynamic nature of the VE requires a continuous verification of its runtime behavior across its entire lifecycle. Existing literature of DTs focusses only on the V&V of the visible behavior of DTs. Muctadir et al.’s [ 20 ] interview research on understanding the current practices involved in the development and maintenance of digital twins found that 13 of the 19 practitioners from both industry and academia were validating their DT by comparing the behavior between VE and AE [ 20 ] . Grieves et al. [ 15 ] discusses visual tests, performance tests, and reflectivity tests for V&V of DTs and these tests also focus on validating the fidelity of the VE. Muñoz et al. [ 21 ] also perform validations of DTs using trace alignments to check for fidelity by comparing the behavior of VE and AE. Dalibor et al. [ 10 ] discuss different techniques such as simulation, testing, model checking and others for validation of DTs. All of these validation methods focus only on validating the DTs at design time and not at run-time, nor do these discuss continuous V&V of DTs. 3 Classical & statistical model checking Formal verification techniques such as CMC ensure complete correctness of a system [ 5 ] . It is a technique of modeling the behavior of a system using a model description language and the requirements of the system are formally specified as properties in order to avoid ambiguity. The model checker performs a complete state space exploration to systematically check whether a specified property holds for (a given state in) that model [ 5 ] . If the property is not satisfied, then most CMC tools provide a counterexample which helps in uncovering issues in the behavior which helps in refining the design of the system. CMC can be applied for verifying reachability properties such as deadlock and livelock freeness, liveness properties, fairness properties and others. However, it suffers from state space explosion issues. Even for simple systems, at times the computed state spaces may become too large for realistic amount of resources. On the other hand, SMC provides probabilistic estimates based