To explore the potential impacts of the chip embargo and gallium-germanium export controls under the backdrop of Sino-U.S. technology rivalry, this paper first clarifies the historical context of Sino-U.S. technological decoupling, then utilizes a theoretical model to illustrate the mechanisms through which these policies affect economic systems. Subsequently, a dynamic computable general equilibrium model is employed to quantitatively assess the economic impacts of the chip embargo and gallium-germanium export controls, complemented by robustness checks. Results indicate that the chip embargo generates substantial negative impacts on China’s economy, while the U.S. economy suffers comparatively smaller losses. China’s gallium-germanium export controls mitigate domestic economic damage, welfare deterioration, and employment reductions caused by the chip embargo, and exacerbate economic losses for embargo-alliance regions, notably Japan, Republic of Korea, and Chinese Taiwan. Furthermore, gallium-germanium export controls weaken the embargo alliance’s competitiveness by constraining their chip manufacturing capabilities, thereby improving the competitive position of China’s electronics industry. This study delineates the economic impact boundaries of the chip embargo and China’s gallium-germanium export control strategies, clarifies the economic trade-offs among different regions, and offers valuable insights for China’s strategic response to U.S. technological containment.
Food security is the cornerstone of national security. In a complex and ever-changing internal and external environment, it is equally important to focus on the resilience and security of the grain industry chain, in addition to paying attention to increasing grain production and efficiency. Based on the perspective of the whole industrial chain, this study constructs an evaluation system for the resilience and security of grain industry chain that takes into account “different stages” and “risk resistance”. Then, it analyzes the spatio-temporal differentiation characteristics and driving mechanisms of the resilience and security of grain industry chain in China’s 31provinces from 2005 to 2021 using exploratory spatial data analysis (ESDA), geographical detectors, geographically and temporally weighted regression (GTWR). Moreover, it deeply explores the impact and spatial spillover effects of the producer subsidy policy using a spatial difference-in-differences model (SDID). It’s found that: 1) From 2005 to 2021, the overall resilience and security of China’s grain industry chain exhibited a steady upward trend, with the production segment consistently serving as the dominant component. The spatial distribution displayed significant positive autocorrelation, while the agglomeration characteristics showed a dynamic evolution trend of“increasing year by year-decreasing fluctuations-tending to stabilize”. 2) The high-value areas in the production stage are mainly concentrated in major grain-producing areas, the high-value areas in the processing stage are more concentrated in the central region, and the high-value areas in the circulation and consumption stages are mostly distributed in developed coastal areas. 3) The explanatory contributions of the various influencing factors, in descending order, are as follows: Land use structure, road network density, annual average temperature, altitude, renewable resource utilization level, urban-rural income gap, digital village construction, and they also exhibited non-linear enhancement and dual-factor enhancement in their interactions, especially the synergistic effect of the renewable resource utilization level is particularly prominent. 4) Further policy impact testing showed the effect of producer subsidy policy has significant temporal and spatial heterogeneity. Its driving effect only lasts for two years, and the optimal policy spillover effect radius is about 400 km.
In the early stage of enterprise development, support through government subsidy rationing is conducive to the cultivation of high-quality enterprises, but the scarcity of government subsidies will trigger the competition for subsidies, and there are incentives for subsidy cheating through surplus manipulation when the enterprise’s qualifications do not meet the subsidy standards. This paper takes A-share listed companies during2011–2021 as a sample, and takes the establishment of national new zones as a quasi-natural experiment to address the strong endogeneity of government subsidies and corporate surplus, and tries to verify whether the control of government subsidy competition will exacerbate corporate surplus manipulation. The empirical results show that enterprises in the new zone are more inclined to negative surplus management to make current profits lower, while the surplus manipulation behavior has a significant double threshold effect based on profits, lower profit enterprises show positive surplus manipulation, higher enterprises show negative surplus manipulation, on the contrary, the profits of the “moderate” enterprises do not have significant surplus manipulation phenomenon. The phenomenon of surplus manipulation is not significant in enterprises with “moderate” profits. Further research finds that the phenomenon of differential surplus manipulation of enterprises has a consistent logic in nature, and all enterprises try to increase government subsidies by smoothing their surpluses through surplus manipulation. In addition, non-state-owned firms, firms with lax financial regulation, and firms with poorer financial endowments are more inclined to compete for government subsidies through surplus manipulation.
This paper uses social network analysis and temporal exponential random graph model to study the structural characteristics and evolution mechanisms of the digital economy correlation network from 2013 to2021. The results show that: 1) The digital economy correlation network has a low density, a short average path length and a large average clustering coefficient. 2) There are four main plates of the network, each with a distinct function. The main beneficiary and net beneficiary plate as the growth pole of the coordinated development of the digital economy, and the effect of “strong combination” among members is obvious, but the effect of “bringing the weak with the strong” is weak. The broker sector composed of Guizhou, Yunnan, Guangxi, Chongqing and Hainan is damaged. The net spillover plate is a supplier of resource, and the number of relationships issued to the outside has decreased during the study period, and it has gradually established two-way links with the rest of the plates. 3) The “group combination” and “broker” structure are crucial to the network’s formation. However, inter-provincial reciprocity is low, resulting in a “ternary closure” characterized by two-way linkages between two of the three provinces, which is not prevalent. 4) The development of the digital economy network is driven by endogenous mechanisms such as preference dependence effect, multi-connection effect, and transmission closure effect. The network is characterized by both path dependence and path creation. In addition, there are differences in the evolution mechanism of digital economy development sub-networks. This study provides new ideas based on the network dependence perspective to promote the coordinated development of China’s regional digital economy.
In recent years, influenced by geopolitical conflicts, resource protection policies, and unexpected events, the global rare earth trade has faced increasingly severe supply risks. This paper constructs a global multi-layer rare earth trade network based on complex network theory and employs a cascading failure model to simulate the propagation patterns of supply risks. The research findings are as follows: 1) The scale, scope, and connectivity of global rare earth trade have shown steady growth. However, due to the global economic environment, the structure of global rare earth trade exhibited a trend toward centralization in 2017 and began transitioning toward decentralization from 2020 onward. 2) China and Australia have gradually become key risk sources in the upstream and midstream networks of the rare earth trade, while European countries (regions) serve as key risk sources in the downstream network. 3) A small number of key countries (regions) control the majority of trade flows and play a dominant role in risk transmission, whereas other countries (regions) have limited impact on others even when subjected to shocks. 4) The risk propagation modes of Australia, China, Germany, and Switzerland are characterized as nuclear propagation, direct propagation, spatial propagation, and indirect propagation, respectively. 5) Rare earth supply disruptions in countries (regions) such as Australia and Vietnam would significantly impact China’s rare earth imports, while supply disruptions in China would substantially affect countries (regions) such as Luxembourg, France, and the United States.
With the frequent occurrence of major events in recent years, urban economies have shown different abilities to withstand, adapt and self-adjust to shocks, which is closely related to the industrial base of cities. As an important manifestation of the industrial base, the impact of industrial collaborative agglomeration on economic resilience is of great significance. From the perspective of synergy between manufacturing and productive service industries, this paper empirically investigates 281 cities’ panel data from 2005 to 2020 on the basis of mechanism analysis, using benchmark regression, mediation and spatial econometric models, and concludes that:Industrial collaborative agglomeration can effectively improve the level of economic resilience, in-depth analysis found that industrial collaborative agglomeration on resistance, resilience and adaptability of the promotion of the more significant, while there is an inhibitory effect on the transformation of the force; in the industry, administrative level of the different cities in the heterogeneity exists. Industrial collaborative agglomeration has a significant positive effect on economic resilience by improving the level of innovation and promoting the upgrading of industrial structure. Industrial cooperative agglomeration has a positive spatial spillover effect on the economic resilience of cities, and an increase in the level of industrial cooperative agglomeration in a city not only enhances its own economic resilience, but also promotes the economic resilience of other cities in the neighbourhood to continue to improve. This paper expands the research on the impact of industrial collaborative agglomeration on economic resilience in China, enriches the empirical evidence of the effect at the city level, and provides useful insights for the city’s industrial layout and the enhancement of economic resilience.
Entrepreneurial spirit is a significant force driving economic development and the academic community has extensively explored its direct impact on economic growth. However, there is a limited body of literature focusing on the diffusion of entrepreneurial spirit among different enterprises. This paper for the first time constructs an independent director entrepreneurial experience database and examines the influence of independent directors’ entrepreneurial experience on the investment efficiency of companies confirming the spillover effects and corresponding mechanisms of entrepreneurial spirit among companies. The research reveals that independent directors with entrepreneurial experience enhance the investment efficiency of companies. After a series of robustness tests such as replacing independent variables, dependent variables, instrumental variables, and changing fixed effects the conclusion remains unchanged. Mechanism tests indicate that entrepreneurial experience empowers these independent directors with a stronger advisory role rather than reinforcing their supervisory effects. Further research shows that the enhancement of corporate investment efficiency by independent directors with entrepreneurial experience is more significant in private firms, when firms are more diversified and when profit growth is faster. This study validates the crucial role of entrepreneurial spirit as a hidden mechanism for economic development. It provides insights into how to better promote and leverage entrepreneurial spirit and serves as empirical evidence for optimizing the independent director system in our country.
This paper incorporates ESG uncertainty and noise into an asset pricing model to explore their effects on the equilibrium prices of assets, investor behavior, and welfare. Our study finds that in scenarios where corporate ESG information is good news, ESG uncertainty increases the proportion of risk assets in ESG-sensitive investors’ portfolios and their welfare while reducing the equilibrium excess returns of risky assets. Conversely, when ESG information is bad news, the opposite effect occurs. The impact of ESG noise on investment strategies and asset prices is distinct from that of ESG uncertainty. ESG noise, which undermines the reliability of all ESG information, leads to more cautious investment behavior among ESG-sensitive investors in good ESG news scenarios, necessitating higher asset premiums for compensation. In bad ESG news scenarios, ESG-sensitive investors tend to take bolder actions, resulting in lower asset premiums. In a market with both green and brown assets, if green assets carry bad ESG news and brown assets good ESG news, ESG uncertainty prompts ESG-sensitive investors to increase their investment in brown assets while reducing it in green assets. When green assets have good ESG news and brown assets have bad ESG news, ESG noise causes ESG-sensitive investors to increase their investment in brown assets and decrease their investment in green assets. Both situations are disadvantageous for green development. This paper affirms the importance of stable green development policies, unified ESG evaluation standards, and regulated ESG rating agency practices in reducing the adverse impacts of ESG uncertainty and noise, offering a valuable addition to domestic research on green development.
With the global energy shortage and environmental pollution becoming more and more serious, how to optimize the energy structure through the construction of effective energy asset portfolio has become a realistic problem that needs to be solved. Most of the existing energy portfolio models only consider economic benefits and ignore their ecological benefits, which limits their applications. For this, this paper explores the impact of factors such as socially responsible investment and carbon intensity on energy portfolio selection, and then constructs a carbon emission reward-penalty mechanism to characterize the portfolio’s ecological benefit. After that, an energy portfolio optimization model with economic and ecological benefits is proposed, where several real constraints including cardinality constraint, budget constraint and risk control demand constraint are taken into consideration. According to the structural features of our model, an improved particle swarm optimization algorithm is designed to solve its optimal portfolio strategy. Subsequently, using real data from the US energy stock market, the paper demonstrates the practicability of the proposed model and explicates the effectiveness of the solution algorithm. Computational results show that the in-sample and out-sample performances of our model are all better than the ones of the existing related models. Namely, investors can obtain an energy portfolio strategy with higher economic and ecological benefits. In addition, the designed algorithm can effectively solve the complex portfolio selection model.
Using A-share listed companies on the Shanghai and Shenzhen stock exchanges during 2007–2022 as the research sample, this paper examines the impact of managerial limited horizon on stock price crash risk and its underlying mechanism. Based on the upper echelons theory and the time orientation theory from social psychology, we construct a managerial limited horizon index using textual analysis and machine learning technology. The empirical results reveal that managerial limited horizon significantly increase the stock price crash risk of the firm. This conclusion remains valid after conducting a series of endogeneity and robustness tests, including two-stage residual inclusion, instrumental variable (IV) analysis, propensity score matching (PSM), etc. The promotion effect of managerial limited horizon on stock price crash risk is primarily achieved through two channels: Disguising loss information and reducing the company’s responsiveness to negative information. Further analysis indicates that the positive impact of managerial limited horizon on stock price crash risk is weakened by non-state ownership, higher management average, larger returnee-manager ratio, higher executive incentives, greater ownership concentration and better external supervision. By integrating managerial limited horizon with stock price crash risk, this study contributes to the existing literature in both areas, which provides empirical evidence for enhancing the quality of listed companies and maintaining the stability of financial market.
Achieving the peak of carbon emissions and carbon neutrality are major environmental and developmental challenges faced by China. In this context, the carbon trading market, as a key policy tool, plays a crucial role in influencing carbon price fluctuations and understanding the prediction mechanisms. Using data from four carbon trading pilot markets in China from 2015 to 2020, the study introduces an innovative method for analyzing the driving factors of carbon trading prices. It employs a long short-term memory network (LSTM) for carbon price modeling and uses SHapley Additive exPlanations (SHAP) values to quantitatively assess the contribution of various influencing factors. This approach enables a comprehensive analysis of overall sample effects, as well as specific factor analysis for individual price fluctuation events. The findings of the study include: 1) The LSTM model exhibits high predictive accuracy, exceeding 90% across all pilot markets, demonstrating its exceptional capability in capturing market dynamics. 2) SHAP value analysis reveals the significant impact of the coal price index, EUA futures prices, the CSI 300 Index, and the average price index of crude oil spot prices on carbon price volatility, indicating a linkage effect between the pilot markets and the EU market. The study also discusses the regional differences in the impact of various influencing factors on carbon trading prices. 3) An analysis based on individual instance samples finds that policy factors are the primary cause of significant fluctuations in carbon prices, exhibiting both positive and negative effects. Based on these findings, this study proposes targeted policy recommendations.
Oil price prediction is a common concern in academia and industry, and good progress has been made in predicting the trend of oil prices. However, in recent years, major crisis events have occurred frequently, leading to frequent and significant changes in crude oil prices, posing new challenges to oil price prediction. This study proposes a research method called “compare real data with predicted data and match influencing factors before making oil prices forecasting” (CRP-MIF-F). Firstly, determine the impact channels of major crisis events within the sample interval. Next, the impact channels of major crisis events are included in the prediction model, and finally, the prediction of crude oil prices is integrated. The prediction results show that the CRP-MIF-F method is significantly better than the benchmark model in trend prediction and turning point prediction, demonstrating good predictive performance This method provides a new perspective for predicting commodity prices considering major crisis events.
In the context of the carbon cap-and-trade mechanism, financially constrained high-carbon emitting enterprises often need to carry out carbon emission reduction technology research and development (R&D)projects through external or internal financing. This approach is utilized to avoid carbon compliance costs. Considering the demand uncertainty and the risk of carbon emission reduction technology R&D, this paper constructs a supply chain emission reduction model under three financing modes: Bank financing, retailer prepayment financing and equity financing. It analyzes the impact of R&D risk on supply chain financing decisions for carbon emission reduction and further explores the choice of financing mode for manufacturers when demand follows a uniform distribution. The study results show that: 1) Under all three financing modes, an increase in the R&D cost coefficient for carbon emission reduction technologies leads to a decrease in manufacturers’ emission reduction efforts and retailers’ product orders. Raising the selling price of products by retailers encourages an increase in their order quantities and enhances manufacturers’ carbon reduction efforts. 2) Under the retailer prepayment financing mode, R&D risk does not directly affect the optimal decision-making of manufacturers and retailers. In contrast, under both bank financing and equity financing modes, retailers’ optimal order quantity and manufacturers’ carbon reduction efforts decrease as R&D risk increases. 3) When demand follows a uniform distribution, the retailer prepayment financing mode is most favorable for promoting carbon reduction among manufacturers and optimizing both the carbon reduction in product production and retailers’ product order quantities.
Based on the uncertain quality of recycling waste power batteries, this study investigates the multi-echelon closed-loop supply chain network design optimization problem with the goal of maximizing system profits. Firstly, a moment-based ambiguity set is proposed to describe the uncertainty of quality, then a two-level distributionally robust optimization (DRO) model is formulated to derive the optimal decision on facility location, recycling pricing and the flow between facilities. Due to the nonlinear complex structure of the model, the dual theory and McCormick linearization method are applied to convert it into an easy-to-solve form, and the Benders decomposition (BD) algorithm is adopted to solve the model. The results show that when the sample size increases, the ambiguity set of the uncertain quality converges to the real data distribution, thereby reducing the conservatism of model and enabling better economic performance in closed-loop supply chain network design. At the same time, compared with stochastic programming, the DRO method proposed in this paper has better robustness in dealing with the uncertainty of recycling quality. Therefore, the closed-loop supply chain network design model constructed in this paper can effectively coordinate the economic and robustness of the waste power battery recycling network design.
Integrating blockchain technology into the fresh food supply chain is a significant approach to addressing consumer distrust regarding the freshness of fresh products. From the dual perspectives of freshness and supplier’s profit, eight game models are established under four preservation strategies: Supplier undertaking preservation, retailer undertaking preservation, third-party logistics service provider (TPL) providing preservation and charging the supplier, and TPL providing preservation and charging the retailer, with and without blockchain adoption. The comparative analysis examines the decisions of fresh food suppliers regarding blockchain adoption and its impacts. The results show that: 1) Under certain conditions, blockchain adoption by suppliers in all four preservation strategies enhances the freshness of fresh products, as well as the profits of suppliers and retailers, albeit with increased wholesale and retail prices of fresh products. 2) Across all four preservation strategies, the threshold cost for suppliers to adopt blockchain considering freshness is higher than when considering supplier profit. When the cost of blockchain adoption is below a certain threshold, it can simultaneously improve both the freshness of fresh products and supplier profits. 3) When the supplier undertakes preservation, TPL provides preservation and charges the supplier, or TPL provides preservation and charges the retailer, there exists a cooperative investment interval among supply chain members for blockchain adoption. In this scenario, the supply chain can achieve a win-win situation through a cost-sharing strategy by the supplier combined with Rubinstein bargaining.
Considering farmers’ capital constraints and yield uncertainty, this study constructs a decision-making model for a green agricultural product supply chain (GAPSC) comprising the government, a bank, an agribusiness, and a farmer. Based on this model, a comparative analysis is conducted of farmer’ production decisions, the company’s level of green technology investment, and government subsidy ratios under different government subsidies. The results are as follows: 1) Government subsidy policies can enhance equilibrium strategies, performance, and social welfare for all parties. When the cost coefficient for green investment is high, subsidies based on purchase volume are more effective for the company. Under both the subsidy scenarios, the green level of the company depends on fairness-concerned parameter of the farmer. 2) Contrary to traditional research results, in the absence of government subsidies, companies do not always adopt a strategy of increasing purchase prices in response to farmers’ fairness concerns, and farmers do not always reduce their planting quantities as a result. When subsidies are based on green technology costs, farmers’ fairness behavior prompts the company to achieve higher levels of green technology investment; however, in the absence of subsidies or with subsidies per unit of purchased volume, farmers’ fairness behavior may negatively impact the level of green technology investment. For banks, the impact of farmers’ fairness concerns on their optimal lending rates shows a trend of first decreasing and then increasing. 3) Except for banks’ optimal rates, yield uncertainty reduces equilibrium decisions, performance, and social welfare for the farmer, company, and the government. Furthermore, as consumers prefer to green agricultural products, most decisions within the agricultural product supply chain exhibit positive changes. This study provides theoretical support for the formulation and selection of green subsidy policies by governments and offers management insights for production and operational decisions in GAPSC.
Customized products play an essential role in market differentiated competition with their unique creativity and personalized characteristics. Many companies have been launching products that are either partially or fully customizable to cater to the increasing demand for personalization from consumers. Building on prior work examining service agents (human employees or robots) and consumer behavior, evidence from five studies demonstrate the effect of service agents (human employees or robots) on consumers’ likelihood of choosing customized products. The research shows that robots enhance consumers’ openness to learning, which leads to higher likelihood of choosing customized products. We also examine a marketer controlled moderator and show that the effect is not observed when the service agent is mechanical robots (vs. thinking robots). The research contributes to both the service agent and the customization literature and offers several practical implications for organizations to effectively choose appropriate service agents.
A virtual enterprise (VE) is an organization that is designed to quickly capture opportunities in the market. Partner selection process is important in the formation phase of a virtual enterprise. This paper studies the process using a bargaining method that follows an alternating offer protocol. Firstly, we consider the complexity of the market in which VEs operate and try to extend bilateral negotiations to one-to-many and many-to-many bargaining. Also we enable agents to have different utility functions that reflect the variability of sellers in terms of price and resource quality, throughout the bargaining process, the ranking of the sellers’ bargaining power is influenced not only by the point in time, but also by the buyer. Then, we use the backward derivation method to obtain the bargaining strategies for agents at every time point under finite time, while considering the market competition, we show that at every time point the agents do not deviate from these strategies, i.e., the bargaining strategies are consistent with the Nash equilibrium. Finally, we use simulation to analyze the process by which agents’ bargaining strategies change dynamically as agents’ relative bargaining power changes.
Building a traceable environment through blockchain to improve the governance level of the online prescription drugs has important practical significance for ensuring the long-term convenience and drug safety of patients with chronic diseases. This study constructs a game-theoretic model between a drug regulatory authority and a drug e-commerce platform, and explores the prescription review and regulatory decision-making and their impact mechanisms. The results show: 1) The patients’ condition, medication compliance, as well as online reviews of drugs constitute the multi-dimensional decision-making framework of the platform and the regulatory authority, while the blockchain significantly changes the review decision-making of the platform by reconstructing the game structure of the platform’s violation cost and income. 2) There is a “double-edged sword effect” in the governance of online prescription drugs enabled by blockchain, which depends on the interaction mechanism between prescription review traceability and online review credibility. 3) Blockchain-enabled prescription traceability regulation mechanism can improve the governance benefits by reducing the regulator cost and the risk of prescription drug abuse, but it may also damage the purchase convenience of patients with chronic diseases and cause social welfare losses.
Evolutionary game theory is an important tool for explaining institutional changes and complex evolutionary phenomena in the field of management. There are significant misuses in existing studies, prominently manifested as the incorrect application of two-dimensional stability discrimination methods in the analysis of high-dimensional systems (such as tripartite evolution). This paper systematically expounds the research paradigm of evolutionary games, that is, to abandon the assumption of complete rationality, focus on the finite rational group in a highly uncertain environment, and through imitation learning and dynamic adjustment, abstract it into a two-dimensional dynamic system to obtain a stable analytical solution, and achieve a stable equilibrium of anti-interference performance under the pressure of the external environment. Aiming at the analytical difficulties of the evolution of high-dimensional systems, four solutions are proposed: Dimension reduction and degradation, numerical simulation, mean field theory, and multi-agent simulation (ABM). This paper clarifies the unique value of evolutionary game theory in revealing the origin and changes of institutions, policy guidance design, path dependence, complexity of policy implementation and prediction of social trends, providing new laws and countermeasures based on adaptive learning for responding to new management challenges under technological changes, drastic environmental changes and the digital revolution.
Compositional data is used to show the proportion of each part in the whole and reflect relative information. Support vector machine (SVM), as a classic classification algorithm, performs well in practical applications. However, the classification of high-dimensional compositional data differs from that of ordinary data, and traditional support vector machines cannot efficiently achieve classification and feature selection. This paper proposes a classification and feature selection algorithm for high-dimensional compositional data based on density convolutional support vector machine (Co-clr-DCSVM). The algorithm is based on the symmetric log-ratio (clr)transformation of compositional data and density convolutional support vector machine (DCSVM), achieving classification and feature selection. In the simulation analysis, the proposed algorithm was compared with other methods, and six evaluation metrics were used to assess the performance of the algorithm. Furthermore, the algorithm was applied to a metabolomics dataset for empirical analysis. The results show that the Co-clr-DCSVM algorithm has high accuracy in classification and can effectively select important features, providing a new solution for dealing with the classification and feature selection of high-dimensional compositional data.
With the promotion of big data and artificial intelligence in the financial credit field, more and more financial institutions are adopting algorithms for automated credit scoring. However, this has led to an increase in algorithmic discrimination. To correct the potential discrimination in automated credit scoring algorithms, this paper builds a fairness learning framework to solve the credit scoring problem of positive and unlabeled(PU) data which commonly appear in the credit market, with the aim of achieving fairness objectives. By classifying the types of discrimination and fairness objectives, corresponding fair PU learning algorithms are developed for each type of scenario, which provides a comprehensive fairness decision-making process for financial institutions to handle PU data and can effectively reduce discrimination in credit scoring while avoiding the estimation bias brought by traditional supervised learning methods. The feasibility of the proposed method and its excellent results in fairness enhancement are verified through simulation experiments and consumer finance example data.
The quality of data is crucial in the age of big data where complete and accurate traffic data is pivotal for cities to achieve an efficient intelligent transportation system. However, in real-world scenarios, traffic flow data often suffer from missing values and outliers caused by sensor failures or transmission errors, significantly affecting system performance. In recent years, low-rank tensor recovery algorithms in the field of machine learning have garnered significant attention due to their ability to capture the intrinsic low-rank structures of high-dimensional data and effectively recover information. However, given that algorithm execution requires inputting batch data, low-rank tensor recovery algorithms typically incur high computational and storage costs when dealing with large-scale data, especially streaming data. Having noticed the quality problem of streaming data prevalent in intelligent transport systems, this paper focuses on the traffic flow data recovery scenario, proposing an online robust tensor recovery algorithm suitable for handling streaming data with side information. The algorithm is validated on a real traffic flow dataset. The results indicate that the proposed algorithm not only achieves better corrupted traffic data recovery by leveraging side information but also exhibits significantly improved execution efficiency compared to traditional low-rank tensor recovery algorithms.
Considering the omission of carbon emissions in conventional vehicle routing problems, this paper incorporates key economic factors such as fuel consumption costs and carbon trading costs into the objective function. It also introduces three-dimensional loading constraints to enhance the model’s comprehensiveness and practical applicability. To solve this complex optimization problem, the paper proposes a two-layer collaborative optimization algorithm. The outer layer uses an adaptive genetic algorithm to optimize vehicle routes, while the inner layer employs a block-loading-based constructive algorithm for validation. This ensures that the resulting routing plans are both theoretically feasible and achieve optimal loading rates. Based on the actual logistics requirements of a logistics enterprise, the paper constructs corresponding instances. It conducts a comparative analysis with ant colony optimization, differential evolution, and discrete particle swarm optimization algorithms. In addition, it performs a vertical evaluation using the standard Augerat benchmark to assess the genetic algorithm before and after improvements. The results validate the applicability of the proposed optimization model and demonstrate the superior performance of the developed algorithm. Furthermore, a sensitivity analysis reveals the driving effect of carbon trading price fluctuations on low-carbon distribution behavior. This research provides strong support for the sustainable development of the logistics and transportation industry.
Path planning is an essential component of unmanned aerial vehicle (UAV) in performing agricultural monitoring tasks. Reasonable path planning methods can improve the operational efficiency of UAV. In order to address the issues of insufficient consideration for multi-UAV scenarios and suboptimal planning effects in current path planning methods, the improved random search algorithm for UAV coverage path planning is studied. Firstly, considering factors such as the number of UAV and its endurance capabilities, a UAV coverage path planning model with the objective of minimizing the total path length is established to obtain efficient UAV coverage path solutions. Secondly, taking into account environmental factors in agricultural monitoring, UAV energy consumption, and flight safety threats, a UAV performance constraint model is constructed. Subsequently, to address environmental uncertainties, the rolling horizon optimization approach is adopted to enable UAV to search for optimal actions within a predicted time window, ensuring the real-time and effectiveness of path planning. Finally, propose the improved random search algorithm and solve the established path planning model and performance constraint model. Simulation results demonstrate that compared to the traditional random search algorithm, the improved algorithm can reduce the total path length and path length discrepancy by up to 10.23% and 99.53% respectively, effectively improving the operational efficiency of UAV, providing reference for research on UAV coverage path planning methods in agricultural monitoring.
