Integrated energy systems (IES) research requires both domain depth and scientific logic, which current large language models lack. To bridge this gap, this study develops EnerAgentic, an IES domain research assistant. Using a Generator-Validator agent pipeline, we automatically built a high-quality Supervised Fine-Tuning (SFT) dataset of approximately 56,000 samples. Based on this, we propose a multi-agent framework that integrates reasoning, retrieval, and tool agents to autonomously handle complex interdisciplinary tasks. Evaluation results show that EnerAgentic comprehensively outperforms open-source models on general benchmarks. Crucially, in the domain-specific evaluation, EnerAgentic-RAG achieved an accuracy of 78.50%, significantly outperforming both its base model (45.00%) and showing highly competitive performance that approaches the GPT-5 (82.50%). This validates EnerAgentic’s core capability to provide end-to-end support for complex data analysis, knowledge retrieval, and simulation modeling in the IES field.

The Dial-a-Ride Problem (DARP) focuses on designing vehicle routes and schedules to transport passengers between specified origins and destinations, with particular relevance to accessible transit systems. This study introduces a novel variant, termed the Heterogeneous DARP with Energy Replenishment Options (H-DARP-ERO), which simultaneously considers heterogeneous user demands with varying resource requirements, a heterogeneous electric vehicle fleet, and two energy replenishment options, namely recharging and battery swapping. A mixed-integer linear programming model is developed to jointly optimize passenger-vehicle assignment, routing, and energy replenishment decisions. To solve the problem, a Population-Guided Hybrid Genetic Algorithm (PG-HGA) is proposed, integrating customized encoding and decoding schemes, local search procedures, and adaptive population regulation mechanism to effectively balance exploration and exploitation. Experiments and a real-world case study demonstrate the efficiency and superiority of the proposed approach, and reveal how replenishment strategies and key parameters affect system performance, providing managerial insights for planning electric heterogeneous Dial-a-Ride systems.

Wind/PV/storage system for communication base station has drawn much attention recently. Considering the insufficient scale effect and poor reliability performance of wind/PV/storage system for a single communication base station at present, the collaborative mode of wind/PV/storage system for communication base station group provides an effective solution. This paper is devoted to optimizing capacity configuration of wind/PV/storage system for communication base station group, considering the uncertainty of wind power, photovoltaic power and load. A multi-objective interval optimization model is constructed with economy, reliability, environmental protection, and social benefits as objective functions, considering the impact of communication load on energy storage charging and discharging. Combining the proposed generalized interval power mean (GIPM) operator with NSGA-II, the interval optimization-NSGA-II-GIPM method is provided to solve the model. Taking the base station group in China as an example, numerical analysis is conducted to verify the effectiveness of the given model and method.

Aborting a mission and initiating emergency rescue represent intuitive strategies for mitigating safety risks during mission execution. This study explores the management of emergency risks through adaptive rescue decisions. Our objective is to develop a dynamic decision-making approach for rescue that minimizes the expected costs associated with system malfunctions and mission failures, while also accounting for mission completion rewards. Most existing mission abort models presuppose a brand-new system prior to mission initiation and focus on formulating abort policies tailored to single-mission scenarios. By contrast, our work incorporates the system's initial age before mission execution and multiple sequential missions. To formalize this complex decision-making process, we conceptualize the sequential abort problem using a Markov decision process (MDP). Our findings reveal that optimal abort decisions rely on establishing control limits, which are derived from key parameters including the distributions of system lifetimes, mission durations, and rescue durations. Additionally, we identify specific conditions that dictate binary decisions: either continuing the mission under certain system states or aborting the mission across all possible states. We further provide sufficient conditions regarding the system's initial age prior to mission execution, which facilitate explicit differentiation between two scenarios: whether initiating the mission is optimal, and whether it is more advisable to avoid starting the mission. To demonstrate the practical relevance of our framework, we present detailed case studies focused on railway signal control systems.

This study examines the stability properties and robustness of nonlinear general type 2 TSK fuzzy control systems (NGT2 TSKs). Due to their higher design flexibility and nonlinear functions applied in their rule consequences, NGT2 TSK fuzzy controllers (NGT2 TSK-Cs) outperform other types of fuzzy controllers in the presence of complex nonlinearities and high degrees of uncertainty. With the help of describing function and parameter plane approaches, the limit cycles and asymptotically stable regions are determined to analyze the stability. Control systems with high unacceptable vibration levels, such as limit cycles, could become unstable as a result. Therefore, particle swarm optimization (PSO) is used to identify the minimum limit cycle region to enhance stability. To assess the robustness of closed-loop systems during uncertainty, minimum gain and phase margins are calculated. Finally, simulation examples are presented to verify the validity of the suggested methodology.

Drones have delivery advantages in some special areas. Green means minimizing the energy consumption of drone delivery. However, minimizing the delivery time or minimizing the delivery distance cannot reflect the delivery cost of drones. Therefore, this paper aims to minimize the total energy consumption. The energy consumption function is established, and the path planning is made for the multi-package delivery of drones. A hybrid algorithm combining heuristic and exact solvers, Advanced Clarke and Wright Saving - Solving Constraint Integer Programs (ACWS-SCIP), is developed to solve the drone delivery problem. In order to improve the speed of solving large-scale problems, another hybrid heuristic algorithm is designed: Advanced Clarke and Wright Saving - Improved crow search algorithm (ACWS-ICSA). In experiments, a case in a remote rural area is studied, and results confirm that the shortest delivery path does not necessarily consume the lowest energy.

This paper investigates the stability of the nonlinear discrete-time strict-feedback system (NDSS) with Poisson noise. This work examines two n-dimensional systems: one with Poisson noise in the n-th state equation, and the other deals with Poisson noise in all state equations. First, a controller is designed for the discrete-time stochastic systems using the backstepping method. Then sufficient conditions are developed to ensure the exponential mean square stability of the system using Lyapunov theory. Finally, numerical examples are presented to demonstrate the usefulness of the proposed results.

Predicting the incubation period of infectious diseases is critical for detecting latent infections. To address this problem, this paper proposes an interpretable machine learning approach called GB-GAMO with gradient boosting acceleration and disjoint region optimization. Gradient boosting acceleration by covariate selection is proposed to speed up the growth of shallow regression trees in training the generalized additive model of individual covariates. Those individual covariates with the largest contribution to loss minimization of greedy function approximation of negative gradients are assigned as the optimal split covariate. Disjoint region optimization is proposed to minimize the loss of residuals in training the generalized additive model on interaction terms. Those interaction terms whose shape functions can minimize the loss of time residuals are used to construct the generalized additive model with optimized weight settings. Experiments on the collected 519 confirmed COVID-19 cases demonstrate that GB-GAMO outperforms state-of-the-art methods in prediction accuracy and interpretability.

In response to the phenomenon of exaggerated publicity in the e-commerce live streaming, many platforms have adopted AI to monitor the live streaming sales process and have taken some targeted measures. In reality, a platform has three punishment strategies to choose from, namely warning, penalty, and pausing. This study investigates the optimal punishment strategy selection for the platform. The results show that regardless of the AI investment level, when the product return rate and brand owner’s exaggerated publicity cost coefficient are high or low, the platform should choose the pausing strategy. When the return rate is low and the cost coefficient is high, or when the return rate is moderate and the cost coefficient is low, the platform should choose the penalty strategy. In the situation of a high AI investment level, the platform should choose the warning strategy when the return rate and cost coefficient are not both high.

The main aim of this paper is to investigate relationship between ideals or filters of a bounded lattice and T-compatible or S-compatible congruences which are obtained by using the definition of compatible aggregation functions. Additionally, we introduce some congruences and characterize the conditions under which they belong to the sets of T-compatible or S-compatible congruences, denoted as Co n T L and Co n S L , respectively. We define an equivalence class denoted by ( a ) I A (A is an aggregation function and I is a nonempty set) and elaborate on the properties of this class when A is t-norm or t-conorm.