A Dataset for Cross-layer Chance-Constrained Optimization of Uncertain Dynamic Systems

To address the complex challenges arising from the interactions among randomness, fuzziness, and cognitive uncertainty in uncertain systems, we first establish a triadic coupling model of random-fuzzy-cognitive uncertainty based on the theory of complete uncertainty spaces. This model achieves the coupling of heterogeneous uncertainty tensors and introduces a comprehensive prediction\u2013correction joint algorithm to overcome time-delay coupling effects. Second, to tackle multi-source uncertainty propagation, we propose a multi-modal uncertainty collaborative quantification framework. By integrating adjoint equation backward derivation, critical propagation paths are identified. Furthermore, the ADMM algorithm is introduced for constraint decomposition in complex tasks. Dynamic weight adjustment enables optimal resource allocation, significantly improving the joint constraint satisfaction rate. Finally, a multi-modal sensor fusion architecture is established to achieve multi-modal fusion. Multi-scenario experiments validate the theoretical and numerical validity of the model and methodology. Experimental results show that the machine vision platform's latency tolerance increases from 18 ms to 35 ms, while the positioning error is reduced to 0.027 mm. Sobol index analysis identifies random noise $S_{\\tau_{i}}=0.78$ coupled with time delay $\\omega_{j}(t)=55 \\%$ as the critical propagation pathway. In industrial robot grasping scenarios, the full-modality fusion approach achieves a 71.4\\% error reduction compared to the monocular vision baseline. Validation in semiconductor defect detection and a 12-inch wafer production line shows significantly reduced defect rates. The multiscale fusion model and dynamic chance-constrained mechanism established in this work provide directly deployable solutions for uncertain systems in smart manufacturing, strongly promoting interdisciplinary convergence in reliability modeling and dynamic control.