This paper is oriented towards investigating the changes in Romanian job market, more specifically focusing on examining the employees' skills in computers and information technology sector. Initially, 12 representative competences (Cognitive skills, Basic technical skills, Advanced technical skills, Software development, Cloud computing, Databases, Data analysis, Machine Learning & Deep Learning, Cybersecurity, Programming engineering, Soft skills, and Personal skills) were selected, and then a Google Forms-based online questionnaire was created and distributed for assessing the respondents' skills with respect to the mentioned categories, involving a Likert scale with 5 points. The case study covers the period between November and December 2024 and consists of 154 valid responses from individuals that are either working or looking for a job in the field of computers and information technology. A grey clustering approach was then used for splitting the respondents into multiple categories, considering their core capabilities. Moreover, the actors in the job market were further described based on the information collected from various sources, and a series of hypotheses were made. Finally, a context-specific ABM framework is proposed, including the agent's types, characteristics, and even possible interconnections, while the use of NetLogo software offers an interface which exposes the identified attributes of the agents under consideration. The insights uncovered in this paper represent essential data for the job market sector, highlighting the most common employees' skills, underdeveloped competences, along with the current employers' distribution and key categories of learning suppliers. This information can assist in further expanding the domain and better comprehending how the technology's evolution influences this area.
The energy consumption system represents a complex, nonlinear, and chaotic system. Starting from multiple nonlinearities within the energy consumption chaotic system. In this paper leverages the characteristics of the Lorenz chaotic system to establish multiple nonlinear differential equations that embody the features of the energy consumption system. Simultaneously, by utilizing the grey differential information inherent in differential and difference equations, a multivariate grey prediction model based on the energy consumption chaotic system is developed. This model not only organically integrates the nonlinear chaoticity of the system with the prediction accuracy of grey prediction models, enabling the simple grey prediction model to reflect the chaotic nature of the energy consumption system, but also meets the adaptive prediction requirements of the energy consumption system and addresses the challenge of predicting multiple energy sources simultaneously. Subsequently, employing the modeling mechanism of the grey prediction model, parameter estimates for the model are obtained. The differential equation set is solved using the discretization method to derive the model's time-response formula, ultimately yielding key modeling steps. To validate the model's effectiveness, the new model is applied to predict the consumption of natural gas, coal, and crude oil in China. Its validity is verified through three types of experiments: the first involves simulation and prediction for different modeling objects, the second effectively tests the model's robustness, and the third compares the new model with other classical grey models. These three types of experiments demonstrate that the new model can effectively capture the complex nonlinear relationships among various variables in the energy system, exhibiting unique stability and significant advantages, along with precise predictive capabilities. It provides robust data support for the planning, scheduling, and management of energy systems.
Accurately forecasting thermal power generation is essential for China's sustainable energy planning and green development. This paper proposes a novel multivariable time-delay grey model to forecast China's thermal power generation from 2025 to 2030. The proposed model combines time-delay effects and dummy variables to capture the complex, nonlinear relationships between thermal power generation and economic drivers under conditions of limited data. By optimizing parameters with the Aquila Optimizer, the proposed model achieves improved adaptability and forecasting accuracy. The proposed model's fitting MAPE is 0.58%, and the test MAPE is 0.29%, outperforming other comparison models. The forecasting results indicate that China's thermal power generation may grow unstably through 2030. The growth rate may slow due to the increasing integration of renewable energy sources and the implementation of carbon neutrality policies. It offers referenceable and foresight insights for the Chinese government to support energy security management and future sustainable development.
Integrating renewable energy into green transportation is critical for achieving sustainable development goals. This study investigates the key factors driving the integration of renewable energy into green transportation systems. Through a comprehensive literature review and expert consultation, 11 drivers were identified and ranked using the Dynamic Grey Relational Analysis (DGRA) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). Government Subsidies and Incentives, Electric Vehicles and Charging Infrastructure, and Energy Security and Reduced Fuel Dependency were the most important drivers. In addition, Kruskal-Wallis test was used to assess the consistency of rankings between different stakeholder groups. Although there were slight differences between the rankings produced by the DGRA and TOPSIS methods, the top five drivers for both methods remained consistent, underscoring their critical importance. The findings highlight that government interventions, such as subsidies and incentives, play a key role in promoting renewable energy adoption. The development of electric vehicles and charging infrastructure, as well as the need to enhance energy security and reduce dependence on fossil fuels, further highlights the systemic nature of these factors. This study provides a robust framework for policymakers and industry stakeholders to prioritize initiatives and efficiently allocate resources to facilitate the integration of renewable energy sources.
Aiming at the problems of scarce data, poor dynamic adaptability of models, and difficulty in uncertainty quantification faced in the prediction of high-temperature creep performance of SiC/SiC ceramic matrix composites for aero-engines, this paper proposes a degradation trajectory model integrating the metabolic GM(1,1) model, stochastic process, and grey cloud theory. By dynamically updating the data sequence, the model captures the creep trend in real time, and uses the grey cloud model to quantify the randomness and fuzziness in the creep process, realizing the cloud droplet distribution characterization of the performance degradation range. Case analysis shows that the proposed model outperforms traditional methods in both prediction accuracy and uncertainty quantification capability, providing an effective tool for the life assessment and reliability design of blade materials under extreme environments.
Attention has emerged as a critical determinant of learning outcomes in an increasingly digital educational landscape. Traditional assessment methods, however, inadequately capture the uncertainty and dynamic fluctuations characteristic of attention data. We developed a Dynamic Nonlinear Grey Relational Analysis (DN-GRA) model integrated with electroencephalography (EEG) neurofeedback technology to enable real-time detection and prediction of attention patterns, thereby facilitating personalized educational interventions. A 12-week randomized controlled trial with 200 primary school students (grades 3-4) revealed that the experimental group significantly outperformed controls across multiple dimensions: attention stability increased 175%, sustained attention duration improved 120%, and average attention values rose 129% (all p < 0.001). Students also demonstrated substantial gains in memory capacity (28.5%), logical reasoning (30.1%), and overall academic performance (effect size d = 0.82). Grey Relational Analysis validation confirmed the effectiveness of personalized intervention strategies, establishing a practical framework for integrating grey system theory (GST) with neurotechnology to advance educational equity and promote sustainable learning outcomes.
In this study, the complex-order damping accumulated GM(1,1) model (CDAGM(z)(1,1)) and the complex-order damping accumulated discrete GM(1,1) model (CDADGM(z)(1,1)) are based on the damping accumulated generating operator and the complex-order accumulated generating operator. By expanding the parameter value range from the real domain to the complex domain, these two new models can better balance the weighting between new and old information, extract more effective features from limited data, and enhance the model accuracy. The relationship between the two models and their respective application scopes was also clarified. An empirical analysis is conducted using China's oil import and export volumes from 2012 to 2024 as the research subject, verifying the applicability and effectiveness of the new models in predicting oil import and export volumes and predicting the oil import and export volumes from 2025 to 2027.
Accurate determination of nursing levels for disabled elderly constitutes a critical issue in long-term care systems. Grounded in grey systems theory, this paper proposes a nursing level discriminative method that integrates nursing task decomposition with discrete grey number operations. Drawing upon Work Breakdown Structure from project management, disabled elderly care is decomposed into eight fundamental nursing tasks. Nursing time data are collected through three consecutive days of observation, discrete grey numbers are constructed, and grey possibility functions are designed for level determination. Case analysis involving five disabled elderly from three nursing institutions successfully assigns each case to the corresponding nursing level. The study reveals that possibility function values for adjacent nursing levels are relatively close for certain elderly, indicating their position in boundary regions and suggesting the need for enhanced monitoring and timely level adjustment. The proposed method effectively handles uncertainty in nursing time, provides definitive level determination while identifying boundary cases, and offers decision support for dynamic monitoring and care resource allocation.
Addressing the nonlinear, fluctuating nature of system data and the insufficient sensitivity of existing grey accumulation to new information, an improved grey Bernoulli model with a damping trend factor (DANGBM(1,N)) is proposed. A damping accumulation operator is introduced, assigning greater weight to new data during preprocessing to enhance sensitivity to recent trends. Combined with the grey Bernoulli model, it better identifies nonlinear characteristics and flexibly adjusts prediction trends, overcoming limitations in existing techniques. Hyperparameters are optimized via particle swarm optimization for adaptability. Applied to China's high-tech industry output value, the model demonstrates superior predictive performance over five benchmarks, offering an effective method for complex nonlinear small-sample forecasting.
The inherent speckle noise in Synthetic Aperture Radar (SAR) images brings difficulty of SAR image segmentation. Especially in the case of high resolution, the real data information of many details is submerged by the speckle noise. A SAR image threshold segmentation method based on grey number and improved flower pollination algorithm (SDEFPA) is proposed in this paper. Segmentation threshold can be regarded as an interval grey number, whitening process can be finished via improved flower pollination algorithm (SDEFPA). In SDEFPA, Sobol sequence strategy is used to initialize the population to improve the population uniformity, and differential evolution mutation strategy is introduced to enhance the ability of flower pollination algorithm (FPA) to escape from local optima. The experiment results on the CEC 2017 benchmark functions show that SDEFPA has better optimization accuracy and convergence speed than 5 comparison algorithms. Application of SDEFPA to SAR image segmentation, grey entropy is selected as the fitness function of SDEFPA. Between global pollination and local pollination randomly switching of multiple cycles by switch probability 0.8, the best flower is gradually close to the optimal threshold. The optimal segmentation is searched quickly by SDEFPA and meanwhile the grey numbers are whitened. Then achieve the final segmented image by whitened threshold. The proposed segmentation method obtains the best effect using the least time compared with contrast segmentation methods. The results demonstrate effectiveness of the proposed method.
Major sudden events have emergency material demands characterized by strong uncertainty, small sample size and temporal complexity. To achieve accurate prediction and scientific scheduling of such material demand, this paper proposes an improved GM(1,1) model. This model is optimized by using Gauss-Legendre integral and GOOSE optimization algorithm, named GLG-GM(1,1). Gauss-Legendre integral is employed to reconstruct both the background value and the time response equation. Meanwhile, to further enhance the model's adaptability, a global adjustment factor is introduced in the background value calculation. With the mean relative error as the optimization objective, the coefficient is globally optimized using the GOOSE optimization algorithm. To verify the model performance, two actual cases are selected to conduct comparative experiments with genetic algorithm GM(1,1) (GAGM(1,1)), discrete fractional-order GM(1,1) (DFOGM(1,1)), Simpson GM(1,1) (SPGM(1,1)), and GM(1,1) models. The performances of these models are objectively evaluated by the mean relative error (MRE), mean square error (MSE), mean absolute error (MAE) and mean absolute percentage error (MAPE). The results show that the prediction accuracy of the GLG-GM(1,1) is significantly better than that of the comparison models, and its adaptability in scenarios with small samples and strong fluctuations is more prominent. The research has confirmed that this model can provide reliable quantitative basis for the planning and dynamic allocation of emergency materials in Hebei Province, China, and provide effective methodological support for improving the regional disaster emergency response capability.
During the operation of an aircraft, the performance of certain systems or components gradually degrades, and both environmental factors and working conditions influence this degradation. Accurately predicting performance degradation by comprehensively accounting for these influences is a critical and complex task aimed at reducing failure-related losses. Moreover, as time progresses, the degradation trend may vary, which imposes higher requirements on the flexibility and adaptability of prediction models. First, this paper introduces the concept of shock effects and models the impact of working condition change using a Beta function while incorporating environmental variables to characterize environmental influences, thereby constructing a grey multivariate prediction model. Second, an optimization model is established to minimize the mean absolute percentage error (MAPE), and the Whale Optimization Algorithm is employed to estimate the unknown parameters related to the shock effects. Subsequently, a recursive least squares-based iterative mechanism is proposed to dynamically and flexibly update the model's structural parameters. Finally, empirical analysis is conducted using hydraulic pressure data from an aircraft hydraulic system under varying working conditions. The results demonstrate that the proposed model exhibits excellent predictive performance and practical value.
The GM(1, 1) power model is frequently employed for forecasting nonlinear data, yet its parameter estimation faces challenges of slow convergence and susceptibility to local optima in existing methods. This study proposes an enhanced GM(1, 1) power model optimized via stochastic gradient descent (SGD), incorporating two innovations: dynamic background-value weights replace fixed ones to adaptively capture data trends, and full-information optimized initial conditions mitigate errors from insufficient information utilization. The SGD algorithm simultaneously optimizes the power exponent, background-value weights, full-information initial condition weights, and model parameters. Comparative experiments with particle swarm optimization (PSO) applied to the same model reveal that SGD, despite marginally slower convergence, achieves significantly higher precision and exhibits superior capability to escape local optima.
This paper first rewrites the first-order accumulated sequence into two distinct sequences, thereby introducing a novel research perspective. Secondly, it reveals the intrinsic relationships and geometric interpretations among the DGM(1,1), NDGM(1,1), QPDGM(1,1), and DGMP(1,1,N) models, and proposes aAnew criterion for determining the degree N in the DGMP(1,1,N) model. Finally, numerical examples and practical applications demonstrate the validity of the proposed theory and the operational feasibility of the method. The approach improves efficiency for applied researchers in model selection and is straightforward to implement.
Accurate forecasting of natural gas production is essential for ensuring energy supply, optimizing the energy structure, and promoting high-quality economic development. This paper proposes a novel Seasonal Weighted Fractional Nonlinear Grey Bernoulli Model (SWFNGBM(1,1|sinx)) to predict China's natural gas production. First, a seasonal weighted fractional accumulation generation operator is introduced and combined with a sine function to enhance the traditional NGBM(1,1) model, aiming to eliminate seasonal fluctuations in the data and capture nonlinear dynamic characteristics. Second, to address the limitations of the traditional Grey Wolf Optimizer (GWO) algorithm in terms of search efficiency and global optimization capability, an improved algorithm incorporating Levy flight is employed to optimize the hyperparameters of the proposed model. Finally, comparative experiments with other models validate the superior predictive accuracy of the proposed model. Based on the newly developed model, China's seasonal natural gas production is forecasted, and the results demonstrate the strong practicality of the model.
The family doctor contract service system represents a multi-actor governance arrangement involving the joint participation of the government, hospitals, family doctors, and patients. this paper develops an integrated analytical framework combining the grey ordinal priority approach (OPA-G) and the graph model for conflict resolution (GMCR) to analyze their strategic conflicts and identify stable equilibrium outcomes. The OPA-G method is first employed to determine the strategic preferences of each stakeholder under uncertain and incomplete information. These preference vectors are then incorporated into the GMCR framework to construct state transitions and assess equilibrium stability under Nash, GMR, SMR, and SEQ criteria. Using the GMCR Plus v0.4 software, the analysis identifies state S1 (YNYYNNYNYNN) as the only strongly stable equilibrium across all four stability definitions. In this state, the government implements strong incentives with strict performance assessments, hospitals enforce internal regulation and evaluation, family doctors actively fulfill contracts and deliver high-quality services, and patients cooperate with service delivery. The proposed Grey OPA-GMCR approach offers methodological guidance for enhancing coordination and optimizing the implementation of family doctor contract services, and more broadly provides a systematic mixed-method framework for modeling multi-actor conflicts in primary healthcare governance.
The limitation of the traditional grey prediction model in handling the derivative order value may lead to unstable or even abnormal prediction results, which suggests that this issue could be a significant influencing factor. In this paper, we unveil a groundbreaking approach to grey prediction by integrating fractional differentiation through a foundation rooted in fractional differential definitions and accumulation principles. The developed model not only allows for dynamic expansion and optimization of both derivative order and accumulation operator within the grey model but also effectively addresses inherent challenges in traditional models. The proposed model exhibits an average minimum MRPE of 2.293%, representing a reduction of 1.936% compared to the average minimum MRPE achieved by the latest comparative model. The result in a more robust and adaptable prediction framework, simultaneously amplifying the model's structural variability and self-accommodating capabilities. This study provides empirical evidence that substantiates the reliability and effectiveness of our innovative prediction algorithm, thereby making a significant contribution to the advancement of the theoretical foundations of grey prediction. Finally, the paper compares the GWO and PSO optimization algorithms. The results show that GWO performs better across all indicators. Therefore, the GWO algorithm is selected to optimize the parameters of the proposed grey prediction model and is applied to three different fields. The prediction results indicate that the average annual growth rates over the next five years will be 1.26%, -1.63%, and -0.23%, respectively.
The time series of nuclear power generation are typically nonlinear, uncertain, and fluctuating data. To solve this, we propose a novel hybrid forecasting framework that integrates wavelet packet decomposition (WPD), a newly developed grey periodic Bernoulli prediction model, and a combined Long Short-Term Memory and Random Forest (LSTM-RF) model. Firstly, decomposing the preprocessed data into low-frequency and high-frequency sublayers using WPD. The low-frequency component is predicted by the novel grey model, whose linear and nonlinear parameters are solved by the least squares method (LSM) and particle swarm optimization (PSO), respectively. Concurrently, the high-frequency component is forecasted by the LSTM-RF model. Thirdly, the forecasting results of the different frequency sublayers are added together to get final prediction result. Finally, the model put into practice to forecast China's nuclear power generation to validate its advantages. The findings indicate that the novel forecasting framework outperforms the benchmark models in terms of performance. This model reduces the MAPE by 51.84%, 33.64%, and 43.53% in comparison with other grey models, and by 32.62%, 31.71%, 24.94%, 48.22%, 44.78%, 32.82%, and 39.54% in comparison with the econometric model and machine learning models.
Accurate prediction of fiber-asphalt bond-slip behavior is crucial for evaluating asphalt concrete performance. Although the generalized grey Verhulst model (GGVM) handles saturated growth sequences well, it exhibits limitations in modeling highly nonlinear interfacial curves due to insufficient "new information priority" in its accumulation operator and crude background value construction that causes error accumulation. To address these issues, this paper proposes a hybrid improved grey generalized Verhulst model (Mar-NIP-FO_GGVM). First, the new information priority fractional-order (NIP-FO) accumulation operator and background value adjustment coefficient are introduced to structurally enhance the model's capability to capture recent information and nonlinear dynamics. Subsequently, particle swarm optimization is employed for global parameter optimization, while the Markov chain is utilized for stochastic correction of prediction residuals. Finally, the model is applied to predict interfacial bond-slip curves of basalt, glass, and polyester fibers. Comparisons with the Popovics model and three other grey models validate the proposed model's effectiveness, applicability, and robustness. Results show that Mar-NIP-FO_GGVM outperforms all benchmarks across three scenarios, with simulation and prediction MAPEs below 7%, while accurately capturing different fibers' stage-wise characteristics. Statistical tests further confirm its significant advantages, establishing a high-precision, data-driven tool for fiber-asphalt micromechanical analysis.
Accurately predicting the remaining useful life (RUL) of lithium-ion batteries is crucial for safe operation. Grey prediction models, with advantages in handling small samples and uncertain information, offer a promising approach for RUL prediction. However, most of the existing grey prediction models focus on the degradation trend while neglecting the capacity regeneration phenomenon. To address this limitation, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) is first applied to separate the capacity degradation trend from local regeneration trends, addressing their differences in magnitude and characteristics. A hybrid prediction method, which combines an improved grey multivariate model and Bayesian-optimized Gaussian process regression, is then proposed. For the capacity degradation trend, which exhibits information heterogeneity and an exponential nonlinear trend, a variable new information priority fractional discrete grey multivariate model is proposed for prediction. The model is not only based on the ideas of variable-order accumulation and discrete grey models, but also introduces an additional nonlinear correction term. For the local regeneration trend, which is nonstationary, nonlinear, and noisy, a denoising autoencoder is employed for noise reduction and feature enrichment, followed by the Bayesian-optimized Gaussian process regression model for prediction. Finally, the predictions of each component are reconstructed to obtain the complete capacity sequence. Multidimensional evaluations on multiple NASA battery datasets, including comparisons with common baseline models and ablation studies to verify the effectiveness of each module, demonstrate that the proposed method achieves superior accuracy, stability, and generalization in capacity degradation prediction.
