In order to reduce the distribution cost of fresh food logistics and achieve the goal of green distribution at the same time, the Green Fresh Food Logistics with Heterogeneous Fleet Vehicle Route Problem (GFLHF-VRP) model is established. Based on the particularity of the model, an improved genetic algorithm called Genetic Algorithm with Adaptive Simulated Annealing Mutation (GAASAM) is proposed in which the mutation operation is upgraded to a simulated annealing mutation operation and its parameters are adjusted by the adaptive operation. The experimental results show that the proposed GAASAM can effectively solve the vehicle routing problem of the proposed model, achieve better performance than the genetic algorithm, and avoid falling into a local optimal trap. The distribution routes obtained by GAASAM are with lower total distribution cost, and achieve the goal of green distribution in which energy, fuel consumption and carbon emissions are reduced at the same time. On the other hand, the proposed GFLHF-VRP and GAASAM can provide a reliable distribution route plan for fresh food logistics enterprises with multiple types of distribution vehicles in real life, which can further reduce the distribution cost and achieve a greener and more environment-friendly distribution solution. The results of this study also provide a managerial method for fresh food logistics enterprises to effectively arrange the distribution work with more social responsibility.

User authentication and verification by gait data based on smartphones’ inertial sensors has gradually attracted increasing attention due to their compact size, portability and affordability. However, the existing approaches often require users to walk on a specific road at a normal walking speed to improve recognition accuracy. In order to recognize gaits under unconstrained conditions on where and how users walk, we proposed a Hybrid Deep Learning Network (HDLN), which combined the advantages of a long short-term memory (LSTM) network and a convolutional neural network (CNN) to reliably extract discriminative features from complex smartphone inertial data. The convergence layer of HDLN was optimized through a spatial pyramid pooling and attention mechanism. The former ensured that the gait features were extracted from more dimensions, and the latter ensured that only important gait information was processed while ignoring unimportant data. Furthermore, we developed an APP that can achieve real-time gait recognition. The experimental results showed that HDLN achieved better performance improvements than CNN, LSTM, DeepConvLSTM and CNN+LSTM by 1.9%, 2.8%, 2.0% and 1.3%, respectively. Furthermore, the experimental results indicated our model’s high scalability and strong suitability in real application scenes.

Real-time tactical sign language recognition enables communication in a silent environment and outside the visual range, and human-computer interaction (HCI) can also be realized. Although the existing methods have high accuracy, they cannot be conveniently implemented in a portable system due to the complexity of their models. In this paper, we present MyoTac, a user-independent real-time tactical sign language classification system that makes the network lightweight through knowledge distillation, so as to balance between high accuracy and execution efficiency. We design tactical convolutional neural networks (TCNN) and bidirectional long short-term memory (B-LSTM) to capture the spatial and temporal features of the signals, respectively, and extract the soft target with knowledge distillation to compress the scale of the neural network by nearly four times without affecting the accuracy. We evaluate MyoTac on 30 tactical sign language (TSL) words based on data from 38 volunteers, including 25 volunteers collecting offline data and 13 volunteers conducting online tests. When dealing with new users, MyoTac achieves an average accuracy of 92.67% and the average recognition time is 2.81 ms. The obtained results show that our approach outperforms other algorithms proposed in the literature, reducing the real-time recognition time by 84.4% with higher accuracy.

With the popularization of logistics technology and the improvement of people’s living standards, the number of chain restaurants with delivery service is increasing continuously due to the convenient service and special discounts. However, the inappropriate management mode and delivery delays often lead to high delivery cost, especially during the peak meal period. Therefore, to solve the problems, two delivery modes, Modern Delivery Mode of Chain Restaurants (MDMCR) and Improved Delivery Mode of Chain Restaurants (IDMCR) were first proposed. IDMCR is an one-stage approach to directly solve the delivery routing problem of multiple branches. Furthermore, we presented an adaptive genetic algorithm for delivery of chain restaurants (AGA-DCR) specifically for IDMCR, which adjusts adaptively the traditional crossover and mutation operations to the fitness of individual and population, therefore, local optimal is avoid. The experiments were conducted with the same dataset to compare different approaches. The results demonstrated that AGA-DCR avoided falling into the local optimal while improved the quality of the solution. Moreover, IDMCR combined with AGA-DCR achieved the best performance, reducing the delivery delay cost by 56.2% compared to MDMCR with GA. Additionally, the optimal delivery routes for chain restaurants were visualized to provide a reference solution for the practical applications.

Accurate forecasting of lottery sales is crucial for strategic planning in volatile consumer markets driven by trend shifts, multi-scale seasonality, and calendar effects. This study proposes a Diagnostic Decomposition–Correction Hybrid (DDC-Hybrid) framework integrating Prophet and SARIMA through a residual diagnostics and correction pipeline. Specifically, Prophet is employed to model long-term trend changes and interpretable holiday impacts, while SARIMA is subsequently used to correct the residual series, capturing short-range temporal dependence that remains statistically significant after decomposition. From an information-theoretic perspective, the framework can be viewed as a two-stage uncertainty reduction process, where decomposition extracts low-frequency informative components and residual correction harvests remaining predictive information. Using monthly lottery sales in China (2008–2025), we conduct a comprehensive evaluation of SARIMA, Prophet, and the proposed hybrid approach. The DDC-Hybrid demonstrates improved predictive accuracy, yielding the lowest error rates. Beyond predictive accuracy, we further examine varying holiday effects through statistical testing. We also find that lottery sales contain a pronounced quadrennial (48-month) seasonal cycle associated with mega-sport events, which improves long-horizon stability. The results suggest that the proposed diagnostic hybrid modeling approach enhances forecasting accuracy and provides practical insights for lottery sales management.

E-commerce retailers bear substantial additional costs arising from high product return rates due to lenient return policies and consumers’ impulsive purchasing. This study aims to accurately predict product return behavior before payment, supporting proactive return management and reducing potential losses. Based on the Graph Transformer, we proposed a novel return prediction model, Returnformer, which focuses on capturing user–product connections represented in topological structures of bipartite graphs. The Returnformer first integrates global topological embeddings into original node features to alleviate structural information loss caused by graph partitioning. It then employs a Graph Transformer to capture long-range user–item dependencies within local subgraphs. In addition, a graph-level attention mechanism is introduced to facilitate the propagation of global return patterns across different subgraphs. Experiments on a real-world e-commerce dataset show that the Returnformer outperforms four machine learning models in terms of prediction accuracy, demonstrating superior performance compared to the state-of-the-art models. The proposed model enables retailers to identify potential return risks prior to payment, thereby supporting timely and proactive preventive interventions.