Consolidating and improving supply chain resilience and maintaining supply chain stability and security is an important foundation for promoting the realization of high-quality development. After initially selecting supply chain resilience related indicators, the research is screened and downgraded through factor analysis to establish a supply chain resilience evaluation index system. Subsequently, based on the model integration framework, the supply chain toughness evaluation model with improved Stacking integration model is constructed on the basis of a single machine learning algorithm and an integrated learning algorithm, and the model parameters are adjusted and optimized through the learning curve to achieve the optimal evaluation effect and compared with the existing model. The results show that the Stacking supply chain toughness evaluation model constructed in this paper has a relative error of 23% or less in 3685 enterprise samples and accounts for 98.78%. It shows that the Stacking integrated model established in this paper has good prediction effect and high accuracy, which has certain value and significance to the research of supply chain toughness prediction, and can provide scientific reference basis for enterprises.

Taxation, as one of the rigid expenditures of enterprises, directly affects the production and manufacturing costs of enterprises, and influences their disposable income as well as the improvement of economic level. Based on the data of Shanghai and Shenzhen A-share listed enterprises in China from 2013 to 2023, the study explores the impact of corporate tax burden on economic development based on multiple regression models, and conducts robustness tests by double-difference and reduced-sample methods. The study finds that corporate tax burden has an inverted “U”-shaped impact on economic development, which is positively correlated with economic development when the corporate tax burden is at a low level, and negatively correlated with economic development when the corporate tax burden crosses the most critical point. In addition, in the heterogeneity test of emerging and non-emerging advantageous industries, the tax burden and the high-quality development of enterprises show an inverted “U”-shaped relationship, but the inflection point of the emerging advantageous industries will appear earlier.

Asphalt as a common material for urban road construction, asphalt pavement has been favored by the municipal industry because of its good wear resistance and stability. This paper identifies the raw materials for this study, and obtains the samples for this study under the technical guidance of the preparation process. In order to more intuitively observe the role of the two, the construction of two-dimensional numerical model, and with the help of water-immersion Marshall test, freeze-thaw splitting test, rutting test, trabecular bending test, together to explore the effect of compounded fibers on the performance of modified asphalt mixtures. When 6% SBR was added on 2% ZnO+4% DTDM, the flexural strength modulus increased from the initial 5482.76 to 6217.54 MPa, and its increase was 18.53%, which indicates that the addition of fiber has a promoting effect on the flexural properties of modified asphalt mixtures.

This study takes the intelligent transportation information management system as the basic framework and focuses on the technical scheme of the traffic flow regulation module in the system. Taking the intersection in urban traffic as the research scenario, we extract the characteristics of urban traffic time and regulation objective function to construct the traffic flow scheduling model. The particle swarm algorithm (PSO) is used to optimize the traffic flow control model, and the inertia weights and the four degree and position update mechanism are improved for the problems of PSO algorithm, such as easy to fall into local optimization. The improved particle swarm algorithm (MPSO) in this paper is utilized to solve the traffic flow scheduling problem, and compared with the PSO algorithm to highlight the effectiveness of the improved operation in this paper. The results show that the optimized traffic flow regulation model based on MPSO algorithm has significant performance advantages in indicators such as average parking delay. Compared with the PSO algorithm, the MPSO algorithm in this paper obviously has higher convergence accuracy and can achieve more excellent regulation solution set in the intersection traffic scheduling scenario. The application of the method in this paper can effectively solve the problems of vehicle congestion and frequent traffic accidents in urban intersections.

In recent years, China’s health insurance drug negotiation has become increasingly improved, and the speed of access to the health insurance catalog has increased dramatically. Under the implementation of health insurance negotiated drugs and dual-channel policy, this study investigates the application of negotiated drugs in a certain region to explore their accessibility and affordability. On this basis, it links the health insurance department, designated pharmacies and medical institutions to explore the precise regulatory mechanism of dual-channel drugs in health insurance. For the drug safety supervision therein, collaborative filtering algorithms, attention mechanisms and multi-task learning are utilized to construct an adverse drug reaction prediction model. It is found that under the influence of the health insurance dual-channel policy, the accessibility and affordability models of medicines are enhanced, the types of negotiated medicines, the number and total amount of purchases are increased year by year, and the total amount of purchases by medical institutions and retail pharmacies are enhanced by 3.42 times and 2.36 times, respectively. The proposed prediction model has good accuracy and applicability in predicting adverse drug reactions, with AUC and AUPR values of 0.93 and 0.83 on different datasets, which are better than the comparison methods. It is recommended to continuously promote the construction of the “dual-channel” management mechanism of designated medical institutions and retail pharmacies to enhance the convenience and sense of access to medical care of the insured. x

The enterprise financial risk under the new era economy exists in all aspects of enterprise operation, thus this paper screens the enterprise financial risk early warning indicators from the four aspects of enterprise profitability, operation ability, development ability, and debt repayment ability. The logistic Steele model is introduced to optimize the population size function of differential evolution algorithm to achieve the dynamic adaptive population size. Then the adaptive differential evolution algorithm is used to optimize the threshold value of BP neural network, and the neural network prediction model based on the improved differential evolution algorithm is derived. Analyze the operation steps of the improved differential neural network algorithm model in enterprise financial security detection to realize the optimal solution of the enterprise financial risk warning model. Compare and analyze the predicted value of the improved differential neural network algorithm model with the real value of the enterprise financial development, and analyze the use of differential evolutionary algorithm in the prediction of enterprise financial risk.The prediction error of the net asset growth rate of enterprise Q in the 1st quarter and the 3rd quarter of the year 2024 is 0.0119 and -0.05309, respectively, with a smaller absolute value of the error, and the improved differential neural network algorithm is able to effectively predict the corporate financial risk.

In recent years, foreign research on the optimal allocation of supply chain resources and operational coordination in the digital economy environment has also made more and more progress, and the current status of domestic research is in the development stage, and supply chain management in the context of the digital economy has become another research hotspot. In this regard, this paper studies the application of Hamiltonian circuit in supply chain resource allocation from three aspects, such as product pricing strategy, supply chain management optimization and consumer behavior, and constructs the optimal resource allocation model according to the steps of resource allocation establishment. Through the Hamiltonian loop algorithm, the revenue function of cross-border ecommerce supply chain services is obtained. Using the Hamiltonian loop algorithm, the optimal price model of supply chain logistics is simulated and simulated experiments are carried out, and the optimal logistics and distribution services will be priced at about 225 under the high competition situation with competition coefficient k  0.6 . Under the optimal allocation of supply chain resources, analyzing the user’s satisfaction, retail customers’ satisfaction with the supply of goods in 2021 is 81.806 points, higher than the province’s 0.913 points, and the experimental results show that the resource allocation model can meet the needs of the customers, making the supply and demand of the product resources more balanced, which argues the scientific nature and reliability of this study.

The accelerated pace of life and social competition become more and more intense, and the problem of psychological pressure faced by people in their study, work and life becomes more and more serious and common, this paper proposes a multi-channel physiological feature fusion method of psychological state assessment for the identification of students’ psychological state in the ideological and political education of college students. The collected multidimensional physiological signal data, such as pulse and picoelectricity, are feature extracted, and the wavelet transform is used to reduce the noise of the physiological signals and realize the waveform filtering, and then the DS evidence theory is combined with the SVM, and the extracted physiological parameters of pulse and picoelectricity are used to realize the effective assessment of psychological stress. Experiments show that the method proposed in this paper of using wavelet decomposition coefficients instead of the original physiological signals as model input can improve the accuracy of psychological stress detection, and the MAPE value of psychological state assessment using the SVM-DS algorithm is 12.28%, which can realize the assessment of students’ psychological state in ideological and political education of college students.

In recent years, art education in colleges and universities has been more and more emphasized by the state and education departments, and has been comprehensively promoted and developed. The study builds the evaluation index system of art education development and assesses the development of art education in a university in order to identify its realistic dilemma. On this basis, the dung beetle algorithm is used to optimize the random forest algorithm to construct a decision tree assessment model of art education development. Through comparison experiments, the prediction accuracy and stability of the DBO-RF model are confirmed, and the deviation of its assessment results from the real value is below 4%, and the RMSE (12.247), MAE (9.133), and MSE (178.829) are lower than that of the comparison method, and the EV (0.721) and R² (0.719) are higher than that of the comparison method, which is applicable to a certain extent. The long-term and overall development of art education in colleges and universities can be promoted by establishing art education mechanisms, strengthening art practice activities, establishing resource sharing channels and developing scientific systems.

In order to improve the design of intelligent products, user cognition and perceptual engineering are integrated into intelligent product design. And through subjective survey and physiological measurements and other techniques to measure the user’s emotional experience of the product, to construct the user’s emotional evaluation model based on BP neural network. Multimodal interaction technology is used to optimize the product design method, and the implicit needs of users for intelligent products are obtained through the method of multimodal perception, which is matched with the product interaction, so as to propose the intelligent product design strategy based on multimodal interaction. In order to verify the effect of the strategy, physiological indicators and perceptual imagery are obtained to evaluate the products. Finally, the user satisfaction of intelligent products under this strategy is studied. The benefit ratio of the smartwatch designed based on the design strategy of this paper (0.438811) is better than other market competitors. The user satisfaction of the 10 experience dimensions of this smartwatch is distributed in the range of [80%, 93%], the dimension with the highest satisfaction is functionality, the lowest is attractiveness, and the overall satisfaction is 86.6%, and the smartwatch designed by this paper’s design strategy obtains a high level of user satisfaction.