Nowadays, smart systems require the use of Digital Twins (DTs) for their engineering and management. Self-updating capability is a key feature in the DT technology. This raises the challenge of model inference from data collected on the system and requires a formal framework be defined, in which a system representation can be coupled with inference methods to achieve automatic model updating. While data-based process model inference is a well-known technique, the inference of a simulation model from existing knowledge and collected data is yet an unexplored area. In this paper, we first clarify and explicitly define some key elements of the terminology related to the DT concept, including model update and model inference, and we propose a framework of inference capabilities based on a formal DT specification, which formally captures the conditions for different updating capabilities and relates them. That way, on one hand, the framework enables the use of symbolic approaches to build a DT with the desired inference capability, and on the other hand, it establishes a partial order relation between inference capabilities. Through a case study, we show how the framework helps formally specifying the DT model of a mobility system toward realizing a fully capable DT.

中文翻译：

---

正式确定数字孪生工程的推理能力框架

如今，智能系统需要使用数字孪生 (DT) 进行工程和管理。自我更新能力是DT技术的一个关键特性。这就提出了从系统上收集的数据进行模型推理的挑战，并且需要定义一个正式的框架，其中系统表示可以与推理方法相结合以实现自动模型更新。虽然基于数据的过程模型推断是一种众所周知的技术，但根据现有知识和收集的数据来推断仿真模型仍然是一个尚未探索的领域。在本文中，我们首先澄清并明确定义了与 DT 概念相关的术语的一些关键要素，包括模型更新和模型推理，并提出了一个基于正式 DT 规范的推理能力框架，该框架正式捕获了不同的更新能力并将它们联系起来。这样，一方面，该框架允许使用符号方法来构建具有所需推理能力的DT，另一方面，它在推理能力之间建立偏序关系。通过案例研究，我们展示了该框架如何帮助正式指定移动系统的 DT 模型，以实现功能齐全的 DT。