Abstract: Driven by the robust growth of ecommerce and the ongoing upgrading of consumer demand structures, the development and strategic layout of private store brands by ecommerce platforms have emerged as a key development trend in the global retail industry. This industrial shift not only grants platforms stronger channel bargaining power and wider profit margins, but also fundamentally reconstructs the traditional channel power structure that was long dominated by national brand manufacturers. Against this backdrop, this study takes a two-tier supply chain consisting of a national brand manufacturer and an ecommerce platform as the research object. Under the channel power structure where the ecommerce platform serves as the leader in the Stackelberg game, we build game models integrating the fairness preference behaviors of supply chain members, to systematically explore the heterogeneous effects of brand advantage, store brand quality perception, and fairness concerns on the pricing strategies and revenue performance of all supply chain participants. The empirical results indicate that in the fairness neutrality scenario, the ecommerce platform’s leadership can significantly boost the revenue of all supply chain members. However, the fairness concerns of the national brand manufacturer are not always conducive to its own benefits, and may even impair the ecommerce platform’s revenue. When the commission rate exceeds a specific critical threshold, overemphasis on fairness may even trigger a drop in the manufacturer’s own revenue. For the ecommerce platform, the revenue contribution of its fairness concern behaviors is heavily dependent on the level of the commission rate, and only exerts a positive promotion effect in the high commission rate scenario. In addition, under the ecommerce platform’s leadership, the increase of the commission rate has a positive driving effect on the retail prices of both national brand and store brand products, while the promotion effect of brand advantage on product quality is restricted by specific parameter thresholds. This study enriches the comparative research on store brand supply chain management under different channel power structures, extends the behavioral operation theory of supply chain members with fairness preference, and offers a new analytical perspective for investigating the core operational decision-making mechanisms of ecommerce supply chains.

Abstract: In order to meet the market demands for customization and rapid response while controlling carbon emissions in manufacturing systems, it has become increasingly important to consider the collaborative optimization of equipment layout in smart production units alongside intelligent storage and transportation systems. This study investigates the loop layout problem of smart manufacturing units equipped with Automated Guided Vehicles (AGVs) through a novel simulation-optimization framework integrating intelligent search algorithms with directed graph-based isomorphism detection. An optimization model is presented with the objective functions of maximizing production capacity and minimizing energy consumption, constrained by performance indicators such as relative position of the equipment within the unit, AGV transport speeds and buffer capacity. Given the lack of a closed mathematical expression for this multi-objective function, a second-generation NSGA-II based on simulation is designed to solve the model. Furthermore, a method based on directed graph isomorphic layout processing is proposed to quickly eliminate similar solutions, enhancing solution quality and algorithm efficiency. Through comparative experimental design and practical applications in intelligent workshops, the effectiveness, superiority, and application value of this optimization algorithm in addressing the circular layout problem of smart production units are validated.

Abstract: Due to the high uncertainty in humanitarian supply chains, utilizing pre-positioning strategies in warehouses to manage crises entails significant risks, including excess inventory and shortages. One approach to managing these risks is to use options contracts. In this study, a dual-option contract, which combines the benefits of both call and put options, is employed to manage risks in a humanitarian supply chain. In this network, in addition to the main supplier, the humanitarian organizations (HO) can also access secondary and tertiary supply sources, namely, in-kind donations and spot markets. Considering these supply sources and the HO's logistics costs including procurement, inventory and transportation, this study models the HO's procurement problem and identifies the optimal initial order, call option, and put option. After forming the centralized problem, the reservation and exercise prices of the dual-option contract are calculated to achieve coordination among supply chain members and a win-win solution. This modeling effort presents a novel application of dual option contracts within humanitarian supply chain, where the contract parameters are endogenously derived to serve as coordination mechanism, an approach not extensively explored in prior literature. Finally, sensitivity analyses are performed on key parameters using numerical examples, and several managerial insights are reported.

Abstract: With the rapid development of the new energy vehicle industry, vehicle-to-grid interaction technology, as an important link connecting the power system and the transportation system, its maturity directly affects the construction of the new power system and the process of energy transition. This paper proposes a maturity assessment method for vehicle-to-network interaction technology that integrates CNN-LSTM and fuzzy comprehensive evaluation. Firstly, a comprehensive evaluation index system was constructed from multiple dimensions such as technology, market, and business model, consisting of five levels: the regulatory layer, the system layer, the market layer, the security protection layer, and the common support layer, covering key indicators such as the accuracy rate of charging load prediction and the return on investment ratio. Secondly, in view of the uncertainty and complexity characteristics in the assessment of technology maturity, the CNN-LSTM model in deep learning was combined with the theory of fuzzy mathematics to construct a dual-path assessment framework. Finally, an empirical analysis was constructed based on multi-source datasets to verify the effectiveness of this method, providing a scientific basis for investment decisions and industrialization promotion of vehicle-to-network interaction technology.

Abstract: Volume-Based Procurement has been implemented as a key reform policy in China's pharmaceutical distribution since 2018, resulting in a sharp decline in the profitability of the pharmaceutical supply chain. This paper develops cost and profit models for Decentralized Decision-Making (DD) and Centralized Decision-Making (CD) in a three-tier supply chain comprising one manufacturer, one distributor, and one hospital. Then, under member profit margin, profit allocation based on the impact of channel power, and the increased profit of each member, a coordination contract is designed that includes the range of coordination factors for the distribution fee rate paid by the manufacturer to the distributor and the subsidies provided by the manufacturer to the hospital. Additionally, the boundary of pharmaceutical pricing under both decisions is investigated, and the findings are finally validated through numerical computation. Some results are found: Under CD, the system’s overall profit increases. Before profit coordination under CD, the logistics costs for the hospital and manufacturer increase, while the distributor’s logistics costs decrease compared with DD, resulting in a concentration of increased profit in the distributor. After coordination, the manufacturer's costs are lower than under DD Furthermore, the pharmaceutical price boundary under CD is lower than that under DD.

Abstract: The present study constructs a remanufacturing supply chain framework that encompasses both manufacturer and retailer. We conduct an in-depth investigation into the influence exerted by blockchain platforms in the supply chain, in response to carbon reduction challenges initiated by manufacturers. By introducing parameters such as the proportion of joint emission reduction investment costs and blockchain unit verification fees, we perform a comprehensive analysis of the effects produced by consumer sensitivity and blockchain platforms on sales prices, recycling prices, carbon emission reduction, market demand, amount of recycled, and profits under different emission reduction modes. Research has revealed that irrespective of whether the manufacturer adopts blockchain platforms, wholesale prices, retail prices, carbon emission reduction, market demand, manufacturer's profit in the joint emission reduction model are always higher than those in individual emission reduction model, while recycling prices and amount of recycled remain unchanged. With retailer's share of the carbon reduction investment cost kept within an appropriate range, both profits of retailer and supply chain increase under joint emission reduction model. When a manufacturer introduces blockchain technology platforms, unit verification fees only affect wholesale prices and have no impact on retail prices, recycling prices, carbon emission reduction, and profits.

Abstract: The prevalence of counterfeit goods hinders the platform economy's growth. To tackle this, we model the strategic interaction between an online platform and a seller using a bilateral evolutionary game. The platform chooses between strict or lenient regulation, while the seller decides on blockchain adoption for product authentication, with government penalties as a key context. We derive the evolutionary stable strategies (ESSs) under four scenarios. Results show that the seller's blockchain adoption depends critically on its cost and the intensity of anti-counterfeiting enforcement by both the government and the platform. Excessively high penalties or insufficient subsidies can deter adoption, highlighting the need for the platform to balance subsidies and penalties. Once the seller adopts blockchain, the platform prefers lenient regulation; otherwise, the platform's decision hinges on her regulatory costs and potential government intervention. Initial strategy choices also significantly impact the evolutionary outcomes.

Abstract: Recently, a growing number of upstream factories are entering the territory of the national brand manufacturer (also called the original equipment manufacturer, denoted by OEM) by establishing their private labels (PLs) in a platform-based supply chain. Yet, existing literature rarely considers how the factory’s entry strategy, non-entry, entry via the platform’s marketplace channel, or entry via the platform’s resale channel, interacts with the OEM’s decision of whether (and how) to introduce a marketplace channel, although this phenomenon is common in reality. In a three-tier supply chain consisting of a factory, an OEM and a platform, we utilize game theory to discuss the relationship between the factory’s entry strategy and the OEM's marketplace channel introduction strategy. Our results show that the factory’s entry via the resale channel always benefits both the platform and himself. By contrast, the factory entering via the marketplace channel may hurt the platform or himself. Furthermore, we show that when the PL's perceived value is low, the OEM’s marketplace channel introduction reduces the probability of the factory’s entry through the marketplace channel; otherwise, such marketplace channel introduction raises the probability that the factory enters the market via the marketplace channel. Surprisingly, we find that the OEM's introduction of the marketplace channel may worsen her profit reduction caused by factory entry. Finally, we derive the equilibrium result and show that in response to factory entry, the OEM chooses not to introduce a marketplace channel to guide the factory to enter the market through the resale channel.

Abstract: To improve the efficiency of mobile robots carrying shelves for material handling in manufacturing enterprises, this study considers multiple material requirements of orders in a coordinated manner and investigates the transfer of multi-load shelves between storage areas and picking workstations. Three types of resources, shelves, automated guided vehicles (AGVs), and storage locations, are jointly considered, and the task allocation and path planning problem for AGVs is studied within an integrated framework. With the objective of minimizing the overall order picking completion time, a mixed-integer programming model is established to simultaneously represent decisions including resource allocation, path selection, and task sequencing, enabling the optimization of multiple interdependent decisions within a unified scheduling framework. To solve the proposed model, an improved genetic algorithm is developed. The algorithm adopts a chromosome encoding scheme based on the "pick–sort–deliver" shelf-handling task unit and constructs multi-granularity crossover operators at the order level, AGV level, and task level to effectively address the multi-resource and multi-level decision structure of the problem. In addition, five neighborhood search operators are designed to form a hierarchical neighborhood search mechanism, further enhancing local solution quality. Experimental results show that the hybrid genetic algorithm significantly outperforms the comparative methods in maximum order completion time, with average improvements of 19.06% and 16.38% over the greedy algorithm and greedy-local search algorithm, respectively. The algorithm also exhibits satisfactory stability and robustness across various problem scales.

Abstract: Effective strategies for Prognostics and Health Management (PHM) are fundamental to advancing the operational integrity and economic sustainability of complex industrial assets. Overcoming the persistent hurdle of modeling the intricate temporal nonlinearities and stochastic behaviors inherent in degradation processes is central to this endeavor. This study introduces an optimized maintenance framework based on real-time condition monitoring for systems subject to stochastic wear, utilizing an innovative nonlinear Wiener process model with time-scale transformation. By framing the optimization objective as the minimization of the long-term average cost rate, we derive explicit expressions for the optimal inspection interval and the threshold for preventive maintenance. A likelihood ratio test demonstrates the superiority of the nonlinear model over linear alternatives. A case study on a Power Take-Off (PTO) unit validates the effectiveness of the proposed method, showing its ability to balance failure risk and maintenance cost efficiently. The proposed strategy provides a practical and scientifically grounded tool for managing systems with nonlinear degradation characteristics.

Abstract: Government trade policies aimed at maximizing economic returns have emerged as a key concern among both scholars and industry stakeholders. Export taxes and subsidies are typically the primary instruments employed. This paper examines optimal export policies for governments within a bilateral trade framework, analyzing both Cournot and Stackelberg competition models. Our model comprises two Manufacturers (Manufacturer 1 in Country A and Manufacturer 2 in Country B), two intermediaries (Intermediary I and II), and a third-Country consumer market. Manufacturer 1 exports to the third market via Intermediary I, while Manufacturer 2 exports via Intermediary II. Crucially, both Manufacturers compete directly within this third market. The analysis reveals that regardless of which export policies governments implement, the social welfare of Countries A and B is influenced not only by the bargaining power between enterprises and intermediaries, but also by the entry order into the competitive third market.

Abstract: Due to the problem of information distortion in the participation of We-Media in food safety supervision, this study establishes a tripartite evolutionary game theoretical framework encompassing government, manufacturers, and We-Media. Then, it employs repeated dynamic equations to examine the system’s equilibrium points and identify evolutionarily stable strategies (ESS). Finally, numerical simulations are conducted. The results show that when We-Media reports demonstrate high credibility, effective media oversight facilitates the evolutionary convergence of the food safety game system toward optimal equilibrium. However, when the authenticity of We-Media reports is weak, society is filled with all kinds of false information and rumors, manufacturers spend a lot of effort to refute rumors, and government public resources are wasted. When government penalties are light under We-Media oversight, manufacturers still choose the strategy of providing low-quality food. Only when the punishment of the government is greater than the profits of the manufacturers do the manufacturers choose to provide high-quality food. In the dynamic system of We-Media's involvement in food safety oversight, there may be three stable equilibrium strategies, among which there are certain connections, which are consistent with the evolution pattern of internet public sentiment. Finally, some management suggestions and countermeasures are put forward.

Abstract: Given the dynamic evolution and institutional relaunch of the global Voluntary Carbon Market (VCM), carbon credit price volatility has emerged as a critical uncertainty influencing the low-carbon transition of non-emission-control enterprises. This paper develops a two-stage stochastic optimization model, integrating heterogeneous risk preferences and carbon option hedging to characterize the joint ordering and emission reduction decisions of non-emissions-control enterprises subject to uncertain carbon credit prices and market demand. The findings indicate that carbon credit price uncertainty significantly diminishes the market participation and emission reduction motivation of risk-averse firms. Carbon options serve as effective hedging tools that stabilize corporate returns while simultaneously strengthening incentives for emission reduction. Notably, carbon price volatility exhibits a dual effect of risk incentive and risk inhibition. Specifically, for risk-neutral firms with hedging capabilities, higher volatility increases option value and encourages more aggressive strategies through a volatility incentive effect. In contrast, risk-averse firms or those without hedging tools perceive volatility as an adverse disturbance that reinforces conservative business behavior.

Abstract: To advance the circular economy, policymakers are increasingly using tools such as carbon trading mechanisms and government subsidies. However, how these policies jointly affect firms’ strategies in such a competitive supply chain setting, a dual-channel closed-loop system featuring both trade-old-for-remanufactured and trade-in schemes, remains unclear. To this end, this study employs a Stackelberg game model to analyze how supply chain actors determine their pricing and recovery strategies when subject to both carbon trading and a dual-subsidy system. Our results reveal a clear divergence between two policy instruments. Dual subsidies consistently expand the market for remanufactured products and increase collected volumes, boosting profits for both firms and enhancing social welfare. In contrast, carbon trading mechanism, while also potentially increasing firm profits, tends to contract remanufacturing market and collection efforts. This research contributes by simultaneously modelling critical and interacting policies. It offers clear guidance for managers on navigating conflicting regulations and for policymakers on designing more effective policy combinations.

Abstract: As technology sanctions increasingly shape contemporary global competition, exemplified by escalating U.S.-China technological decoupling, firms in critical domains confront complex strategic choices between autonomous and imported innovation pathways. While existing research has advanced our understanding of innovation responses to external pressures, significant gaps remain in explaining how corporate innovation strategies dynamically evolve under sanction conditions, particularly regarding the bounded nature of innovation ecosystems where limited actors engage in continuous strategic adaptation amid heightened environmental uncertainty. To address this limitation, we develop a finite-population stochastic evolutionary game model based on the Moran process, systematically analyzing how technology sanctions reshape innovation ecosystems through strategic interactions between autonomous and imported innovation approaches. Using Huawei's comprehensive sanction experience as our empirical foundation, we conduct multi-context simulations across three distinct scenarios: expectation-driven contexts (where rational payoff comparisons dominate), stochastic-dominant contexts (where random perturbations significantly influence decisions), and dual-uncertainty contexts (where both payoff ambiguity and environmental volatility coexist). Our analysis reveals that technology sanctions drive innovation strategy evolution through progressively complex mechanisms: initially operating via dual constraints of potential losses and market erosion, then expanding into quadruple interactions involving policy support and domestic market effects, and ultimately evolving into dynamic regulatory systems with fluctuation-bounded characteristics. Crucially, we demonstrate that while autonomous innovation emerges as evolutionarily stable under the first two contexts, dual-uncertainty scenarios preclude stable equilibria entirely, leading to persistent strategic coexistence. These findings provide theoretical foundations for understanding how institutional shocks reshape innovation ecosystems and offer practical guidance for corporate strategy formulation under technological fragmentation.

Abstract: This paper investigates an integrated production and distribution scheduling problem in a hybrid flow shop with a batch delivery system. The goal is to schedule jobs and assign them to batches to minimize the total weighted number of tardy jobs and delivery costs, thereby improving both operational efficiency and customer satisfaction. To achieve this, a novel mixed-integer linear programming model is developed. Given its computational complexity for large instances, two metaheuristic algorithms, simulated annealing (SA) and a genetic algorithm (GA), are proposed and calibrated using the Taguchi method to enhance performance. Experimental results demonstrate that simulated annealing consistently outperforms the genetic algorithm in both solution quality and computational time. The results also demonstrate that the proposed approach reduces the overall cost of production and distribution by 10.45%.

Abstract: Minimizing the internal supply chains’ operating expenses is a crucial management goal in current transnational enterprises, where manufacturing and retailing often operate independently. Still, managers must periodically evaluate the consolidated operating expenses and performance. The operational goals in the manufacturing units consist of lowering quality- and reliability-relevant costs, avoiding manufacturing delays due to random breakdowns, and meeting order due dates through expediting strategies, such as partial subcontracting and accelerating the fabrication plan. Inspired by efforts to optimize batch runtime for the mentioned intra-supply chains, this study investigates the combined impact of quality-surety actions (including defect removal and rework), correction of breakdowns, multi-delivery, adjustable-rate, and outsourcing on a coordinated producer–retailer system. This study presents a research scheme comprising: (1) model development for the mentioned internal supply-chain features; and (2) optimization approaches for determining the cycle time decision that minimizes the overall system operating expenses. To conclude our work, we validate the research scheme, procedure, and results through numerical demonstration and show that it can effectively support management’s decision-making with various exploratory and crucial information.

Abstract: To address the critical issues of inefficient coordination and suboptimal global performance in Heterogeneous Robot Order Fulfillment Systems (HROFS), this study investigates the Heterogeneous Robot Pick-up and Delivery (HRPD) problem by mapping it as an Open Shop Scheduling problem with Sequence-dependent Setup, Transportation Time, and Transportation Waiting Time (OSS-SSTTWT). A multi-objective scheduling model is introduced to enhance collaboration between Autonomous Mobile Robots (AMRs) and picking robots by simultaneously minimizing Total Completion Time, transportation time, and transportation waiting time. To achieve Heterogeneous Robot Adaptive Collaborative Task Planning (HR-ACTP), an Adaptive Multi-Objective Deep Q-Network improved with an Upper Confidence Bound strategy (UCB-AMDQN) is proposed. Experimental results demonstrate that the UCB-AMDQN significantly outperforms industrial baselines such as FIFO, as well as DDQN and combined scheduling rules, by discovering synergistic collaboration patterns that transcend human-designed heuristics and existing cognitive frameworks. Furthermore, the research analyzes the impact of velocity differences and partitioning strategies on coordination efficiency. By explicitly incorporating often-neglected transportation metrics into a novel OSS-SSTTWT framework, this work provides a comprehensive and effective pathway for optimizing complex heterogeneous robot collaborative task planning in smart logistics.

Abstract: This paper constructs a dual-channel Stackelberg game model including the benchmark wholesale model B, the manufacturer encroachment model E, and the front-end and back-end AI-enabled model A. Through comparative static analysis, it is found that the manufacturer's encroachment profit increases with the increase of consumers' green preferences and channel substitutability. When the effective market capacity is positive, the manufacturer's profit under the encroachment strategy is better than that of the benchmark model, and the retailer's profit under the encroachment model is not lower than that under the benchmark model. Additionally, the optimal AI investment increases monotonically with the effective market capacity and efficiency parameters and decreases monotonically with the cost coefficient. When reaching a certain threshold, it shows a significant upward trend. Moreover, the stronger the channel substitutability, the lower the threshold. Under given constraints, the profits of both the manufacturer and the retailer in model A are significantly higher than those in other models. Therefore, moderate front-end and back-end AI collaboration is the key to achieving the triple goals of "encroachment + green + digitalization" and achieving a win-win profit situation between the manufacturer and the retailer. This provides a theoretical basis for enterprises to design direct sales strategies, AI investment intensity, and for governments and platforms to formulate subsidy and cost-sharing policies.

Abstract: The growing prevalence of product idling weakens strategic consumers’ immediate purchase willingness, while the rise of peer to peer (P2P) secondary platforms creates value enhancement effects for consumers and cannibalization effects for manufacturers. Against this backdrop, this study constructs a two-period dynamic model that incorporates used product transactions to analyze a manufacturer’s pricing strategy for new product launches and to examine how the manufacturer performance, consumer surplus, and social welfare are affected by the probability of product idling and the P2P platform’s availability. The findings demonstrate that the manufacturer should adopt a skimming pricing strategy, with price levels modulated by the product idling probability. Although price reductions can be effective as idle probability rises, increasing later-period prices may also improve profitability. The prospect of idling further delays consumer purchases, redirecting subsequent demand toward used (new) products when idle probability is low (high). While rising idle rates typically lower manufacturer profit and social welfare, consumers may benefit under certain conditions. Furthermore, in a market with a P2P platform, the manufacturer balances the inhibitive effect of idling anticipation against the platform’s dual effects by adjusting prices according to idle probability: lowering (raising) both current and future prices if it is low (high), and raising current while lowering future prices if it is moderate. Although the platform consistently reduces manufacturer profit, it could enhance consumer surplus and social welfare. Finally, extending the analysis to scenarios with imperfect matching between used-product supply and demand further corroborates the robustness of these findings.

Abstract: As the integrated energy system continues to deepen towards multi-energy complementarity and collaborative optimization, the coupling and interaction among its internal electric, thermal, gas, storage and other multi-energy networks have become increasingly complex. The emergent behaviors triggered by the nonlinear interactions among intelligent agents have further increased the uncertainty of system state prediction. In response to the shortcomings of traditional prediction models, such as insufficient generalization ability, difficulty in adapting to cross-scenario dynamic changes, and ignoring the influence of emergent behaviors, this paper proposes a meta-adaptive learning framework that integrates the perception of emergent behaviors of intelligent agents, aiming to build a high-precision state prediction model for the integrated energy system. Firstly, a multi-agent interaction structure and emergent behavior modeling structure for the integrated energy system are designed to quantify the emergent features generated by the coupling and interaction among intelligent agents; Secondly, a meta-adaptive learning core structure is constructed, where the meta-learner extracts cross-scenario general knowledge and combines the adaptive modulation mechanism of the basic learner to achieve dynamic scene adaptation; Finally, a state prediction execution structure is designed to complete feature fusion and precise prediction. Experimental results show that the proposed model achieves an average absolute error of 0.023 on three typical scenario datasets, a root mean square error of 0.031, and an average absolute percentage error within 1.8%. Compared with traditional LSTM, Transformer, and ordinary meta-learning models, the prediction accuracy is improved by 25% to 42%; in cross-scenario transfer tasks, the adaptation time of the model is shortened by 68%, and the computing efficiency is increased by more than 35%; meanwhile, in the scenario of sudden load fluctuations, the model still maintains prediction stability, with the average absolute percentage error fluctuation not exceeding 0.3%, providing efficient and reliable technical support for real-time scheduling and safe operation of the integrated energy system, and applicable to the state prediction scenarios of complex multi-energy coupling and dynamic scene switching of the integrated energy system.

Abstract: This paper addresses the Permutation Flow Shop Scheduling Problem with Heterogeneous Workers (PFSP-HW), an extension of the classical problem in which processing times depend not only on the job and machine, but also on the assigned worker. This variant better reflects practical environments where worker capabilities and proficiencies vary significantly. We propose a new Iterated Greedy (IG) heuristic adapted to handle worker heterogeneity. The IG heuristic combines destruction and reconstruction mechanisms with a local search procedure tailored for the problem. We develop two versions of the proposed algorithm and compare them with adapted state-of-the-art heuristics and metaheuristics from related problems. The algorithms were tested on a large benchmark set comprising 360 instances generated under various shop configurations. The suggested IG heuristics surpass current approaches in terms of solution quality and execution time, as determined by computational and statistical evaluations, making them reliable and efficient tools for solving the PFSP-HW.

Abstract: This study proposes a dynamic inventory optimization model based on deep reinforcement learning (DRL) to improve the intelligence of inventory management in multi-node supply chain systems, aiming to mitigate stockout risks and reduce inventory holding costs. The model is built upon three key components. First, a simulation environment is developed to mirror real-world supply chain operations, incorporating multiple products, warehouses, and sales channels. The simulator takes into account critical factors such as lead times, demand variability, and order fulfillment constraints. Second, the Soft Actor-Critic (SAC) algorithm is employed as the core of the learning strategy. To enhance the model’s adaptability to dynamic inventory changes, a dual-stage state representation and an attention-enhanced perception mechanism are introduced. Third, the model is trained and validated using the publicly available Instacart Market Basket Analysis dataset, which serves as a benchmark platform for testing retail replenishment strategies. To assess its effectiveness, the proposed SAC-based model is compared with several baseline methods, including Deep Deterministic Policy Gradient (DDPG), Proximal Policy Optimization (PPO), and an LSTM-based Model Predictive Control approach (LSTM-MPC). Experimental results across six representative scenarios show that the SAC model consistently outperforms competing approaches. Specifically, in high-demand volatility and limited-capacity scenarios, the SAC model reduces average inventory costs to ¥69,900 and ¥97,200, respectively, substantially lower than those achieved by DDPG (¥120,700 and ¥136,300). Under compound disturbance conditions, SAC limits the number of stockout events to 59, which is 50% fewer than the rule-based benchmark, significantly improving service level performance. Moreover, in high-frequency, short-cycle environments, SAC achieves the lowest inventory variability, with a standard deviation of just 4.12, indicating superior policy stability. Overall, the proposed DRL-based model exhibits strong robustness and adaptability, offering a practical solution for intelligent and resilient inventory control. This study highlights the potential of advanced reinforcement learning methods in addressing complex scheduling problems in real-world supply chain management.

Abstract: The flight process of unmanned aerial vehicles (UAVs) is influenced by various uncertain parameters, such as air density, aerodynamic coefficients, and propulsion system thrust. Prior to mission execution, predicting the flight envelope under the impact of uncertainties is critical, and uncertainty quantification (UQ) serves as the key method to address this challenge. Traditionally, in UAV UQ analysis, unknown parameters are often assigned specific distributions based on prior knowledge. However, such priors may be subjective and are frequently set overly conservatively to meet safety margins. Conversely, historical flight data holds immense value for quantifying uncertainties in reusable or identical-model UAVs. This paper proposes a method that utilizes historical flight data to estimate parameters through Bayesian inference, then fuses these estimates with prior knowledge to establish more objective distributions for uncertain parameters. Reasonable parameter distributions positively impact UQ by preventing control strategies from being insufficiently robust or overly redundant. Therefore, this paper explores uncertainty quantification for UAVs under different information sources and accelerates the algorithm using Gaussian Process Regression (GPR) and Polynomial Chaotic Expansion (PCE).

Abstract: In the context of the digital economy, especially with the application of blockchain in recent years, it has had a profound impact on the pharmaceutical manufacturing supply chain. Based on this, this study focuses on the strategic choices of channel governance models for medical manufacturing enterprises and has constructed three pharmaceutical supply chain production models: the benchmark wholesale price model S, the profit-cost sharing coordination model C, and the blockchain occupation model E. The research found that the coordination model C has the greatest advantage when cost sharing is sufficient and profit distribution is reasonable, which can effectively motivate channel members to expand the market scale and ultimately achieve a win-win situation. The benchmark model S plays an important coordinating role when the market risk is high. The occupation model E can significantly enhance the market dominance when the enterprise has strong blockchain technology capabilities, but it will also cause channel conflict risks. In addition, this study provides theoretical basis for achieving a positive synergy loop in terms of market prices, services, blockchain investment, and sales volume, and provides strong guidance for the strategic management choices of supply chain members.

Abstract: With the diversification trend of stock market data, constructing a stock price prediction model that can integrate multi-source data, handle high-dimensional features, and fit the complex relationships between predictive and response variables is crucial for formulating effective quantitative trading strategies. Based on this, this paper proposes a stock price prediction model that combines multi-kernel learning with Gaussian Process Neural Networks. The main innovations of the model are reflected in the following aspects: First, by using kernel functions with different domains, it effectively integrates feature information from various data sources; Second, by combining Gaussian Process Regression with neural networks, the model fully considers the impact of features from different data sources on the prediction results while addressing the curse of dimensionality and maintaining good fitting ability for complex relationships. In addition, the model can quantify the uncertainty of stock price prediction results and perform statistical inference analysis, providing more detailed auxiliary information for investors or decision-makers. Simulation analysis results show that the proposed method outperforms some classic models in predictive performance, showing strong competitiveness. Finally, in quantitative trading back-testing, considering the introduction of non-Euclidean sentiment features, the model has achieved significant superior performance in indicators such as cumulative returns, Sharpe ratio, and maximum drawdown.

Abstract: Amid increasingly stringent environmental regulations and Chinese “dual-carbon” goals, the manufacturing sector faces simultaneous pressures of economic performance and environmental responsibility. Remanufacturing, with its advantages in energy savings, emissions reduction, and resource reuse, has become a key pathway to green manufacturing; however, market growth is hampered by consumers’ distrust of product quality and information opacity. This study incorporates potential manufacturer encroachment and introduces blockchain technology into the remanufacturing market. We build a two-echelon supply chain model consisting of a manufacturer and a retailer, and develop three variants: a baseline model, a blockchain cost-sharing model, and a manufacturer-encroachment model. Using a Stackelberg game framework, we analyze the optimal decisions and profit outcomes of the manufacturer and the retailer under different blockchain cost levels, degrees of consumer awareness, and encroachment intensities. The results show that blockchain enhances consumer trust by improving information transparency, but its economic benefits depend critically on implementation costs and consumers’ understanding of product information. When encroachment intensity is moderate, both the manufacturer and the retailer can achieve a win–win outcome, with higher overall supply-chain efficiency and increased sales of remanufactured products. Further analysis reveals a misalignment between the manufacturer’s profit-maximizing point and the environmentally optimal outcome; this can be reconciled through piecewise government subsidies. This paper enriches research at the intersection of blockchain and green supply-chain governance and provides theoretical guidance and practical implications for firms formulating remanufacturing strategies and for policymakers designing green incentive mechanisms.

Abstract: This study addresses the daily scheduling and routing problem for home healthcare workers while incorporating lunch break requirements. The Home Healthcare Scheduling and Routing Problem is analysed alongside its common constraints, including patient and caregiver time windows, caregiver qualifications, and mandated breaks. To address this, four different variants of an effective Adaptive Large Neighbourhood Search (ALNS) algorithm were developed to provide high-quality solutions. The algorithms demonstrate significant efficiency, solving 30-patient instances optimally within an average of 12 seconds. For scenarios involving 100 patients, they maintained robust performance with a slight increase in computational time of about 54 seconds. Results indicate operational efficiency improvements of up to 36% through optimized travel routes and patient visitation schedules. To translate these findings into practice, a decision support system, the Home Healthcare Decision Support System (HHDSS), was designed to assist administrators by automating the complex task of scheduling and routing of caregivers. Tested using realistic patient data generated from Turkey, the system effectively allocates healthcare resources and improves responsiveness. Overall, the proposed framework shows strong potential as a valuable practical tool for improving the responsiveness and efficiency of home healthcare logistics.