Sustainable Diffusion-based Incentive Mechanism for Generative AI-driven Digital Twins in Industrial Cyber-Physical Systems

TL;DR

This work addresses the challenge of building and maintaining GenAI-driven Digital Twins in ICPSs when IIoT data sharing is hampered by information asymmetry. It combines a GenAI-driven DT architecture with a contract-theory incentive design and a sustainable diffusion-based Soft Actor-Critic method that uses dynamic structured pruning to compute optimal contracts efficiently ($K$ types of IIoT devices, IR/IC constraints). The approach demonstrates improved DT construction and updates while reducing training carbon emissions, highlighting practical scalability for intelligent manufacturing. The findings offer a pathway to energy-efficient, incentive-aligned data sharing that enhances predictive maintenance and real-time decision-making in industrial ecosystems.

Abstract

Industrial Cyber-Physical Systems (ICPSs) are an integral component of modern manufacturing and industries. By digitizing data throughout product life cycles, Digital Twins (DTs) in ICPSs enable a shift from current industrial infrastructures to intelligent and adaptive infrastructures. Thanks to data process capability, Generative Artificial Intelligence (GenAI) can drive the construction and update of DTs to improve predictive accuracy and prepare for diverse smart manufacturing. However, mechanisms that leverage Industrial Internet of Things (IIoT) devices to share sensing data for DT construction are susceptible to adverse selection problems. In this paper, we first develop a GenAI-driven DT architecture in ICPSs. To address the adverse selection problem caused by information asymmetry, we propose a contract theory model and develop a sustainable diffusion-based soft actor-critic algorithm to identify the optimal feasible contract. Specifically, we leverage dynamic structured pruning techniques to reduce parameter numbers of actor networks, allowing sustainability and efficient implementation of the proposed algorithm. Numerical results demonstrate the effectiveness of the proposed scheme and the algorithm, enabling efficient DT construction and updates to monitor and manage ICPSs.

Figures (7)

• Figure 1: The architecture of GenAI-driven DTs in ICPSs. We study how GenAI drives the DT construction pipeline, i.e., real-time physical data collection, communications for DTs, DT modeling and maintenance, and DT decision-making.
• Figure 2: The proposed algorithm architecture, where we utilize dynamic structured pruning techniques to sparsify the actor networks of the diffusion model. The goal of the algorithm is to design optimal contracts that motivate IIoT devices to provide sensing industrial data, which is one of the data sources for DT construction, as shown in Fig. \ref{['Digital']}.
• Figure 3: Test reward comparison of the proposed scheme with the random scheme under asymmetric information and the contract theory scheme under complete information.
• Figure 4: Performance comparison of the proposed algorithm with several DRL algorithms in optimal contract design. For the parameter settings of the proposed algorithm, we set the pruning rate to $10\%$, the diffusion step to $6$, the learning rate of actor networks to $2\times10^{-7}$, and the learning rate of critic networks to $2\times10^{-6}$.
• Figure 5: The utility of the DT server and the average utility of IIoT devices under different algorithms.

Theorems & Definitions (3)

• Definition 1
• Definition 2
• Remark 1