For the insufficient prediction accuracy and systematic bias in traditional single machine learning model for elective surgery scheduling, this study innovatively proposes a multi-model fusion framework based on the grey possibility function, which improves prediction accuracy through a grey possibility-function-based representative fusion mechanism. A grey representative prediction system was constructed by integrating machine learning algorithms such as gradient boosting, random forest, and decision trees. Notably, this system introduces an innovative three-layer surgical duration classification framework that autonomously adjusts prediction parameters based on the characteristics of surgical durations. This feature effectively addresses the traditional model's tendency to underestimate long surgeries and overestimate shorter ones. In the verification phase, multidimensional evaluation is conducted using prediction accuracy indicators and scheduling efficiency indicators. Experimental data shows that the new framework achieves a 0.13% reduction in MAE indicators and a 3.0% increase in R-2. In terms of surgical scheduling efficiency, the average total overlap time was optimized to 8.81 minutes, a decrease of 36.48%. More importantly, the prediction bias of long and short surgeries narrowed to-14.2% and +45.4%. This research provides reliable data support for the allocation of operating room resources. By accurately predicting the surgery duration, it can improve the utilization rate of operating rooms, optimize the scheduling of medical staff, and rationalize the scheduling of medical equipment, ultimately promoting the overall improvement of medical service quality.
Selecting an enterprise resource planning (ERP) system that simultaneously satisfies sustainability, circular economy, and Industry 4.0 requirements is a multi-criteria decision problem complicated by incomplete information and small expert panels, in which judgments are more naturally expressed as ordinal priorities than as precise scores. To address the lack of an integrated sustainable-ERP (S-ERP) selection framework for these conditions, this study develops and applies a Grey System Theory (GST)-based Grey Ordinal Priority Approach (OPA-G). The framework embeds sustainability, circular economy, and Industry 4.0 criteria in a single decision structure and represents the judgments of a three-member expert panel as ordinal priorities under grey interval uncertainty, avoiding pairwise comparisons while preserving uncertainty through grey numbers. The model is applied to a real manufacturing case, and its robustness is tested through seven expert-priority scenarios together with area-level (+/- 10%, +/- 20%) and ordinal perturbation analyses, and benchmarked against Grey TOPSIS and Grey VIKOR. The results show that the Fundamental and Economic criteria dominate the decision (0.640), followed by Industry 4.0 compatibility (0.233), environmental (0.099), and social (0.028) criteria; System Cost, Quality, and Financial Situation are the most influential criteria, while Cloud and Edge Computing, Artificial Intelligence, and IoT are the leading digital requirements. The framework yields the ranking ERP1 > ERP2 > ERP3 > ERP4, which remains unchanged across all perturbation scenarios and is reproduced identically by Grey TOPSIS and Grey VIKOR. Although economic performance remains primary, Industry 4.0 capabilities are the decisive enablers that make environmental and social monitoring operationally executable. The proposed framework therefore offers a transparent, robust, and practically implementable decision-support tool for sustainable ERP procurement under uncertainty, extensible to multi-firm panels and broader sustainability domains.
To address the issues in existing panel data-based grey relational analysis models (PD-GRA), such as the neglect of indicator's time delay and data transformation considerations, a time-delay grey relational analysis model (TDGRA) is proposed. First, the concepts of the sample matrix, the submatrix under time delay, and the corresponding truncation matrix are introduced. Second, singular value analysis is performed on the relevant matrices, and the optimal rotation matrix is derived. Then, an optimization model is constructed with the goal of minimizing the Procrustes distance, through which the time delays between indicators are determined. And then, a PD-GRA model is proposed based on the optimal Procrustes distance with time delay. Finally, the TDGRA model is used to identify the driving factors of AQI in the Yellow River Basin, obtaining urbanization rate (UR), carbon emissions (CE), gross domestic product (GDP), and the growth of total investments in fixed assets (GIFA) as the main drivers. The results demonstrate the rationality and effectiveness of the proposed TDGRA model, and its advantages are further highlighted through comparative analysis.
The grey power model captures nonlinear trends, but its accuracy heavily depends on the power exponent, which is often empirically selected. For instance, the grey Verhulst model and GM(1,1), two special cases of grey power models and widely applied, struggle to adapt to the fitting of complex data. To address this, we propose an Improved Grey Power Model (IGPM) that optimizes the exponent gamma via a dual-path optimization strategy integrating grid search and the Artificial Fish Swarm Algorithm (AFSA) within the theoretically determined bounds of [-2, 2]. In the context of small-sample data, validated on annual power generation and metal fatigue strength, the optimized IGPM reduces MAPE by 37.5% and RMSE by 38.0% on average over the standard GM(1,1). The framework provides a robust solution for data-scarce engineering predictions.
Preference learning is one of the important methods in machine learning and has been applied in various fields. Preference learning in the presence of uncertain information is still a field that deserves further attention. The main objective of this paper is to construct a preference learning model for three-parameter interval grey numbers (TPIGNs) using the UTilit & eacute;s Additives (UTA) method, to solve two key issues in preference learning: (1) failing to find the optimal preference, and (2) having multiple optimal preferences complicates the selection process. To address the two issues, this paper proposes a preference-learning method based on the UTA model for TPIGNs considering the inconsistent decision maker's pairwise preferences. TPIGNs are first converted into whitening values under a triangular possibility distribution and then incorporated into an additive value-function model. The first-rank acceptability index from stochastic multicriteria acceptability analysis (SMAA) is used to measure the prior importance of each reference alternative, from which an importance index is defined for each pairwise preference constraint. An improved UTA model is then constructed to identify and remove low-importance contradictory preferences through binary variables while retaining more informative preference statements. A random-direction extreme-point sampling procedure is used to obtain representative value-function parameters and improve the robustness of the ranking result. A supplier-selection case involving age-friendly equipment illustrates the implementation of the method and its advantage over the traditional UTA method in handling inconsistent preference information.
