What is happening worldwide in digital twins, mathematical modelling and the management of large-scale systems: research, press releases, product news. From the US, China, ASEAN and Africa — translated and summarised.
300 items · 100 more in the archive
Wireless digital twins (DTs) are increasingly integrated with artificial intelligence to support online modeling, prediction, and network optimization. However, most existing approaches perform adaptation within predefined learning or optimization structures, which may become unsuitable when…
Schneider Electric, AVEVA, Nscale and Start Campus pilot AI digital twin Engineering.com
arXiv:2610.28907v1 Announce Type: cross Abstract: Multi-agent reinforcement learning (MARL) offers a principled way of populating simulated economies with many heterogeneous decision makers.
arXiv:2610.29559v1 Announce Type: new Abstract: Model predictive control is a high-performance control method, but it incurs comparatively high online computational costs across many control tasks and is challenging to realize in low-level hardware settings.
arXiv:2610.29377v1 Announce Type: cross Abstract: Learning-based path planners based on diffusion and flow matching can generate diverse trajectories from demonstrations but typically lack mechanisms for constraint satisfaction and collision avoidance during deployment.
arXiv:2604.04455v3 Announce Type: replace Abstract: Nonlinear underactuated systems such as two-wheeled inverted pendulums (TWIPs) exhibit a limited region of attraction (RoA), which defines the set of initial conditions from which the closed-loop system converges to the equilibrium.
Collective behaviors such as synchronization transitions are emergent properties of coupled oscillator systems, and how to reproduce them from data for large-scale systems remains an open problem.
A digital twin (DT) for type 1 diabetes (T1D) can predict how an individual's blood glucose will respond to different insulin treatments without actually trying them. For each treatment, downstream statistics of predicted glucose trajectories, such as glycemic risk, can be computed to guide…
Vivix Executable Digital Twin for Float Glass Snowflake
Metal additive manufacturing (AM) offers substantial advantages to the fabrication of complex structural components, but its widespread adoption is slowed due to complexities in the process-structure-property relations which is the cause for uncertainty and drives risk.
The same TAdviser item: the generative AI market reached 58 billion rubles in 2025, with 44 billion corporate. Measurable results are becoming the key criterion, while most projects remain pilots due to integration, data quality, security and cost issues.
Russia's generative AI market reached 58 billion rubles in 2025, with 44 billion in the corporate segment. Spending growth has not translated into comparable results: many projects remain pilots, and moving to operation depends on integration, data quality, security and processing cost.
The work addresses the execution gap in industrial cells, where commands may take effect after the plant state has changed. The Action Intent Schema lets a producer declare validity and assumptions, while the XAIR runtime rechecks each intent and releases it as a version-conditional command executed only if the state version is unchanged.
Using a synthetic population of one million agents in an urban setting, the study compares five mobility models differing in activity patterns and destination choice. It shows how mobility simplifications affect emergent contact networks and epidemic outcomes.
The paper reviews current uses of LLMs in agent-based modelling and critically examines opportunities and challenges across the modelling cycle, from problem formulation to documentation and communication of results, including summarization, information extraction and reasoning.
The DA-KCF framework keeps an offline-identified nominal quadrotor model calibrated with limited hardware data and learns only a lightweight online correction for disturbances from causal state-command-state transitions. These corrections are evaluated along the prediction horizon inside MPC to optimize position and yaw references.
Gasification of Bishkek residential areas will continue from 2028, according to the Kyrgyz parliament speaker citing Gazprom's local subsidiary. Completion of a pipeline from Russia is named as the condition.
Conventional surgical segmentation relies on closed-set categories or explicit text queries, which cannot resolve implicit queries that require reasoning about workflows, instrument functions, and temporal relationships.
The increasing complexity of Industrial Control Systems (ICS) and their integration with IT infrastructures has significantly expanded the cyber threat landscape. Traditional cybersecurity models, primarily based on the Confidentiality, Integrity, and Availability (CIA triad), fail to capture the…
Inline quality assurance for Sheet Molding Compound (SMC) manufacturing is hindered by the lack of large-scale, accurately labeled image datasets for training machine-learning models. This work presents a process-informed, photorealistic digital twin for carbon-fiber SMC strip deposition.
Sber's applied AI center and ITMO University have set up a laboratory focused on multi-agent AI technologies. The source gives no detail on scope or specific tasks.
A framework is proposed for stochastic model predictive control of nonlinear systems with parametric uncertainty and time-varying disturbances. The Koopman operator is parameterized by polynomial chaos expansions, combining data-driven dynamics with uncertainty propagation. The resulting deterministic reformulation keeps the online optimization dimension dependent only on control horizon and input dimension.
A distributed iterative method is presented for cooperative linear MPC with polyhedral state and input constraints, using a logarithmic barrier function. The centralized problem is turned into a distributed scheme with a convex nonlinear solver guaranteeing global linear convergence. A fixed iteration count per sampling step ensures closed-loop stability at constant computational cost.
A hybrid model predictive control approach treats switching instants as explicit optimization variables instead of relying on a fixed time grid. Each step solves a single sparse quadratic program jointly optimizing inputs and impact timing. Optimality conditions are consistent with the hybrid minimum principle.
