Urban traffic congestion leads to frequent vehicle start-stop events and low-speed operation, which is one of the primary drivers of carbon emission growth. To address the problems of multi-objective conflict, training instability, and inadequate carbon emission modeling in existing traffic signal control methods for carbon emission optimization, this paper proposes a deep reinforcement learning signal control algorithm for carbon emission optimization. This method constructs a carbon-emission-aware dynamic reward mechanism and achieves collaborative optimization of traffic efficiency and emission reduction objectives through adaptive weight adjustment; Lagrange multiplier method is introduced to embed the carbon emission threshold as an explicit constraint into the strategy learning process to ensure that the emission level is controlled within an acceptable range; For multi-intersection scenarios, a distributed collaborative control framework based on parameter sharing and neighborhood information interaction is designed to enhance the model's ability to perceive the spatial propagation characteristics of traffic flow. Based on the SUMO simulation platform, experimental validation is conducted in three scenarios: a single intersection, a 4×4 grid network, and a real-world urban road network. The results show that compared with PPO algorithm, the average carbon emissions of this method are reduced by 11.3% to 12.8%, average delay is reduced by 15.7%, average speed is increased by 9.6%, and the comprehensive performance index is improved by 12.2%; During the training process, the fluctuation of strategy is reduced by about 50%, and the degradation rate of generalization performance is reduced by 34.2% compared with the comparison method. This study provides an effective intelligent solution for low-carbon-oriented urban traffic signal control.

Aiming at the problems of strong subjectivity, lack of accuracy and difficulty in large-scale evaluation of students' sports action standardization, this paper proposes an automatic evaluation algorithm based on computer vision. First, a multi-perspective sports action dataset is constructed and an expert scoring system is designed; Secondly, key point sequences are extracted using an improved pose estimation model, and a multi-scale motion representation method is introduced to integrate joint-level, limb-level, and global features; Furthermore, a bias-aware alignment network is proposed to achieve adaptive modeling of spatiotemporal errors; Finally, a multi-task scoring model based on the fusion of GCN and Transformer is constructed to realize the normative classification and regression prediction of actions. The experimental results show that on the self-built data set, the MAE of this method is reduced to 0.318, which represents an improvement of approximately 29.6% over mainstream methods, the classification accuracy is 91.6%, and the correlation coefficient with expert score is 0.94. At the same time, in the cross-scenario test, the performance decreased by only 2.8%, which was significantly better than the comparison method. Ablation experiments and statistical tests validate the effectiveness of each module. The results show that this method has obvious advantages in accuracy, generalization ability and interpretability, and can provide technical support for intelligent physical education teaching and automatic evaluation.

The Urban Growth Boundary (UGB) is essential for controlling urban sprawl and optimizing land use, yet traditional delineation methods struggle with modeling complex spatial relationships and fusing multi-source data. This study proposes an intelligent UGB delineation algorithm based on a Graph Neural Network (GNN). The study area is discretized into uniform spatial units to construct an urban graph, with node features integrating remote sensing imagery, land use types, transportation networks, and population-economic data. A spatially constrained graph convolution structure with an improved attention mechanism is designed to jointly model spatial structures and expansion driving factors, enhanced by multi-scale feature aggregation and spatial consistency constraints. Experimental validation in a 1,250 km² urban area (5,024 nodes, approximately 3.8×10⁴ edges) demonstrates that the proposed model achieves 0.912 accuracy and 0.900 F1-score in UGB recognition—3.4% and 3.8% higher than traditional GCN—with a 7.2% improvement in spatial consistency. The model remains stable across 250–1000 m spatial scales, indicating strong generalization ability and spatial adaptability. This GNN-based UGB delineation method effectively captures urban spatial structure characteristics and expansion patterns, providing a high-precision, data-driven technical pathway for territorial spatial planning and sustainable urban growth management.

As a key indicator of macroeconomic performance, inflation trends significantly influence monetary policy, macroeconomic regulation, and financial market stability. However, macroeconomic time series often contain both linear trends and complex nonlinear fluctuations, which limit the accuracy and stability of traditional statistical models. To address this, the paper proposes an improved ARIMA-LSTM hybrid forecasting model for inflation rate prediction. The ARIMA component extracts the linear structure of the series, while the residual sequence captures unexplained nonlinear information. A multi-scale LSTM network then learns deep features from the residuals, and a dynamic weight fusion mechanism adaptively combines linear and nonlinear predictions. Experiments using CPI data from IMF, World Bank, and FRED databases show that the proposed model achieves an RMSE of 0.564 on the test set—12.1% lower than the traditional ARIMA-LSTM and 17.1% lower than ARIMA alone. It also outperforms models such as SVR, random forest, LSTM, and GRU in MAE and MAPE. In multi-step forecasting, error growth remains around 12% over six steps, notably lower than comparison models. Ablation studies and Diebold–Mariano tests confirm the effectiveness of the multi-scale module and dynamic fusion mechanism. Overall, the improved ARIMA-LSTM model enhances inflation prediction accuracy and stability, offering practical value for macroeconomic forecasting and policy analysis.

Because cross-border trade fraud involves multiple types of entities, multiple business relationships and complex interactive structures, it exhibits high heterogeneity and strong concealment, which has brought significant challenges to the traditional risk identification methods. Aiming at the problem that existing methods struggle to balance the ability of structural modeling and classification performance, this paper proposes a cross-border trade fraud detection framework based on heterogeneous graph neural network (HGNN) and gradient lifting tree model XGBoost. Firstly, the cross-border trade system is modeled as a heterogeneous graph of multi type entities and multi relationship interactions, and HGNN is used to learn the high-order structural semantic representation of entities in complex trade networks; Then, the graph embedding features and statistical features are input into XGBoost to achieve high-precision classification of fraud. The experimental results on the real cross-border trade data set show that the AUC of the proposed model on the test set reaches 0.966, which is 18.7% and 3.4% higher than using XGBoost and HGNN alone, and significantly improves the recall rate of fraud samples in a variety of typical fraud scenarios. Ablation experiments further verified the key role of heterogeneous relationship modeling, attention mechanism and integration strategy in performance improvement. The above results show that HGNN–XGBoost integration framework has good detection performance and engineering application potential in complex heterogeneous scenes.

Based on the resource-based view and dynamic capability theory, this study constructs a chain-mediated model to examine the impact of supply chain digital integration on the operational resilience of manufacturing enterprises. Through empirical analysis of 480 pairs of manufacturing enterprises and logistics service providers in seven Central and Eastern European countries, the study finds that: First, supply chain digital integration significantly improves enterprise operational resilience ( , ). Second, digital integration indirectly enhances operational resilience through two pathways: operational synergy (indirect effect = 0.197, 95% CI [0.152, 0.254]) and knowledge synergy (indirect effect = 0.170, 95% CI [0.128, 0.226]), with mediating effects accounting for 34.1% and 26.4% of the total effect, respectively. Furthermore, relationship stability (interaction term , ) and technological capability matching (interaction term , ) both significantly enhanced the role of digital integration in promoting resilience. Heterogeneity analysis further indicated that this effect was more pronounced in regions with weaker institutional environments, SMEs, and highly complex manufacturing industries. This study expands the theoretical framework of supply chain digitalization and resilience from a bilateral collaborative perspective and provides empirical evidence for manufacturing enterprises in Central and Eastern Europe to build systemic resilience through digital integration.

The short-term export volume forecast is of great significance for international trade decision-making and macroeconomic regulation, but the time series of export volume usually contains significant seasonal, trend and nonlinear fluctuation characteristics at the same time, so it is difficult to obtain ideal results with a single forecast model. In order to improve the prediction accuracy and stability, this paper proposes a hybrid prediction method combining seasonal autoregressive moving average model (SARIMA) and attention based memory network (attention based LSTM). Firstly, SARIMA model is used to describe the linear structure and seasonal components in the export volume series, and its prediction residual is modeled; Then, the attention LSTM is used to learn the nonlinear dynamic characteristics in the residual sequence, and finally the prediction results are obtained by additive fusion. The experimental results based on the monthly export volume data of Poland show that compared with the traditional SARIMA, LSTM and other comparative models, the MAE, RMSE and MAPE of the SARIMA-Attention-LSTM (Proposed) on the test set are reduced, on average, by about 20%–35%, respectively. The prediction residual fluctuation converges significantly, and shows better stability and generalization ability in repeated experiments. The results show that the effective integration of statistical model and deep learning model can significantly improve the short-term export forecasting performance, and provide a feasible and efficient solution for the prediction of complex economic time series.

This paper addresses the urgent needs of intelligent manufacturing for highly reliable equipment operation and precise maintenance and delves into the innovative application of digital twin technology in production line condition monitoring and predictive maintenance. By systematically reviewing the core theories of digital twins, multi-source heterogeneous data acquisition and processing technologies, and predictive maintenance methodologies, a five-dimensional integrated framework comprising a physical layer, data layer, model layer, functional layer, and application layer is constructed. This framework innovatively achieves real-time dynamic mapping and bidirectional interaction between physical and virtual spaces, establishes a data-model hybrid-driven mechanism for equipment health status assessment and remaining life prediction, and forms a closed-loop optimization system from condition perception and fault early warning to maintenance decision-making. To verify the effectiveness of the framework, this study conducts a case study using a precision CNC gear machining production line. The results show that the framework can control the latency of critical equipment condition monitoring within 200 milliseconds, improve the accuracy of remaining life prediction by approximately 15% compared to purely data-driven methods, successfully achieve early fault warning, reduce unplanned downtime by 65%, and save maintenance costs by 28%. The research findings provide theoretical guidance and practical examples for achieving precise and forward-looking equipment health management in the context of intelligent manufacturing and have important reference value for promoting the digital transformation of the manufacturing industry.

This paper, grounded in the contemporary imperative of "cultivating virtue and nurturing talent," explores strategies to integrate Chengdu's rich revolutionary cultural heritage into practical teaching of ideological and political theory courses at local universities. The study first analyzes the distinctive connotations and educational value of Chengdu's revolutionary culture. It then systematically examines three primary implementation models: the "classroom narrative + resource immersion" approach, the "on-site teaching + emotional resonance" model, and the "project-driven + knowledge-action integration" framework. Through reflective analysis of practical outcomes, the paper identifies challenges in resource integration, content adaptation, and evaluation mechanisms. It proposes optimization recommendations including establishing a collaborative "university-local-library" education mechanism, developing digital teaching resources, and refining formative assessment systems. These insights aim to provide actionable strategies for enhancing the appeal, engagement, and effectiveness of ideological and political education in regional universities.

With the acceleration of urbanization and the deepening of the concept of sustainable development, miniaturized personal transportation vehicles have become a key carrier for solving the "last mile" travel demand. However, current market products generally face the dilemma of homogenized styling and a lack of emotional value. Meanwhile, biomimetic design, as an interdisciplinary innovation method, provides a new perspective for breaking through this bottleneck. This study aims to systematically explore the application value and practical path of biomimetic design methods in the styling of miniaturized personal transportation vehicles. Through literature review, case analysis, and morphological deconstruction, it deeply analyzes the internal logic and manifestations of three application modes: morphological biomimetic, functional biomimetic, and imagery biomimetic. The study finds that biomimetic design can endow products with natural beauty and vitality through organic forms, optimize aerodynamic performance and lightweight structure with biological prototypes, and evoke emotional resonance in users through abstract imagery. The study further proposes an integrated strategy of "form-function-emotion," emphasizing that the selection of biological prototypes should be deeply aligned with user cognition and usage scenarios. Despite the challenges of engineering implementation and over-mimicry, biomimetic design combining intelligent interaction and sustainable technologies will become an important evolutionary direction for future personalized mobility experiences. This study provides a theoretical basis and practical inspiration for constructing miniaturized transportation vehicles that combine aesthetic value, functional rationality, and emotional warmth.