This paper introduces Graphon Attachment Network Models (GAN-M), a novel framework for modeling evolving networks with rich structural dependencies, grounded in graphon theory. GAN-M provides a flexible and interpretable foundation for studying network formation by leveraging graphon functions to define attachment probabilities, thereby combining the strengths of graphons with a temporal perspective. A key contribution of this work is a methodology for learning structural changes in these networks over time. Our approach uses graph counts—frequencies of substructures such as triangles and stars—to capture shifts in network topology. We propose a new statistic designed to learn changes in the resulting piecewise polynomial signals and develop an efficient method for change detection, supported by theoretical guarantees. Numerical experiments demonstrate the effectiveness of our approach across various network settings, highlighting its potential for dynamic network analysis.

Graphons offer a powerful framework for modeling large-scale networks, yet estimation remains challenging. We propose a novel approach that leverages a low-rank additive representation, yielding both a low-rank connection probability matrix and a low-rank graphon--two goals rarely achieved jointly. Our method resolves identification issues and enables an efficient sequential algorithm based on subgraph counts and interpolation. We establish consistency and demonstrate strong empirical performance in terms of computational efficiency and estimation accuracy through simulations and data analysis.

This work presents the first theoretically justified simultaneous inference framework for off-policy evaluation (OPE). In contrast to existing methods that focus on point estimates or pointwise confidence intervals (CIs), the new framework quantifies global uncertainty across an infinite or continuous initial state space, offering valid inference over the entire state space. Our method leverages sieve-based Q-function estimation and (high-dimensional) Gaussian approximation techniques over convex regions, which further motivates a new multiplier bootstrap algorithm for constructing asymptotically correct simultaneous confidence regions (SCRs). The widths of the SCRs exceed those of the pointwise CIs by only a logarithmic factor, indicating that our procedure is nearly optimal in terms of efficiency. The effectiveness of the proposed approach is demonstrated through simulations and analysis of the OhioT1DM dataset.

With the development of battery technology, large-scale battery applications are increasing. In order to obtain a higher current and voltage level and improve the overall energy efficiency, batteries are connected in series and parallel. Bulk model is the most used model to simulate battery packs, and the simulation results of single cell are enlarged several times to represent a battery pack. But bulk model ignores inconsistency between the batteries, and when the current magnification is large, such as in the extreme fast charge scenario, this neglect cannot reflect the current distribution inside the battery pack. Our paper proposes a simplify algorithm based on equivalent circuit model, which applied for non-uniform series and parallel-connected batteries. This simplification can greatly improve simulation speed while maintaining accuracy under the consideration of inconsistency. More importantly, this algorithm applies to any complex circuit, regardless of how many series-parallel structure layers. With the proposed algorithm, Monte Carlo simulations were evaluated to obtain the influence of battery series and parallel number and normal distribution parameters on current distribution. Simulation results show that an increase in the number of series will reduce the inconsistency of the current distribution, and an increase in the number of parallel and an increase in the relative standard deviation of the normal distribution of the battery parameters will increase the inconsistency of the current distribution.

With the development of extreme fast charging technology, charging stations need to use energy storage stations to reduce the rising peak to average power ratio (PAPR). Lithium-ion capacitor (LIC) is a chemical power source that uses both Faraday process and non-Faraday process to store energy. Because of its attractive performance in terms of rate characteristics and chemical stability, it is suitable for some energy storage stations that consider both power density and energy density. It is important to describe the current-voltage characteristics of LIC to predict the charge and discharge efficiency in the early design of energy storage power stations. During the test, however, a full discharge or charge results in a high temperature rise, and the electrical model parameters near a specific temperature point cannot be accurately obtained. The short current pulses cannot stabilize the polarization. In this paper, a high-accuracy parameters identification method based on an improved Butler-Volmer-Equation-Based electrical model is used to summarize the phenomena caused by the rate of change of high-energy LIC. The accuracy of the method is tested under the dynamic stress condition test. The maximum voltage error is less than 2%. Energy efficiency calculation based on the used model is simulated by the design condition from the energy storage station of the Haizhu line in Guangzhou. The maximum error is less than 0.2%.