As outsourcing evolves from a temporary solution into an entrenched practice, efficiently scheduling production under critical operation constraints has become a pivotal challenge.Thus,this paper proposes a flexible job shop scheduling problem with critical operation driven outsourcing strategy and interval grey processing time (G-FJSPOC). The strategy is based on a pre-defined classification where critical operations are mandatory for in-house processing as a hard constraint, while non-critical operations can be outsourced based on real-time machine utilization. This approach ensures security and quality for critical operations while introducing flexibility for non-critical ones. To characterize temporal uncertainty in the manufacturing process, We employ interval grey numbers to describe processing times and refine their operational rules. A multi-objective model is established with makespan, total cost, and dynamic carbon emissions as optimization targets. For the proposed G-FJSPOC, an enhanced imperialist competitive algorithm with embedded simulated annealing (EICA-ESA) is developed. First, Logistic-Tent dual chaotic mapping generates the initial population to enhance solution diversity and quality. Subsequently, the simulated annealing algorithm was incorporated into the evolutionary phase, achieving a balance between local development and global exploration. Furthermore, a multi-strategy competition mechanism is applied to continuously optimize the population structure and enhance the algorithm’s exploration and convergence capabilities. The hybrid design enables effective exploration of the multi-modal solution space under uncertainty and operational constraints. Experimental analysis through 20 test cases demonstrates the effectiveness of the proposed EICA-ESA for solving G-FJSPOC.
Urban Rail Transit (URT) station areas are pivotal nodes for shaping urban structure and fostering Urban Vibrancy. However, the spatio-temporal heterogeneity of the built environment’s impact on urban vibrancy remains underexplored, as existing research predominantly relies on spatially invariant global models, which overlook how the effects of the built environment vary across geographical locations and temporal contexts. To address this gap, this study uses Xi’an Metro Line 2 in China as a case study. Employing multi-source big data, including Location-Based Service (LBS) population heatmaps and Points of Interest (POIs) at a fine-grained grid scale, we utilize a Geographically Weighted Regression (GWR) model to quantitatively deconstruct the spatially varying interactions between built environment’s variables and urban vibrancy across different time periods (morning, noon, and evening).The results first statistically confirm the significant spatial non-stationarity of the built environment’s impacts. Based on this, the study identifies three distinct vibrancy generation archetypes: (1) "endogenous service-oriented" vibrancy, driven by localized built capacity and public services; (2) "exogenous attraction-based" vibrancy, defined by city-level functional synergies and public space quality; and (3) "nodal absorption-oriented" vibrancy, characterized by the efficient absorption of commuter flows through high-density development and functional diversity.These findings empirically challenge the universal applicability of principles within Transit-Oriented Development (TOD) theory. The core contribution of this study, therefore, is its advocacy for a paradigm shift in planning-from a "prescriptive framework" that applies standardized principles to a "diagnostic framework" predicated on the precise identification of local, dominant driving mechanisms. This conclusion provides a solid scientific basis and theoretical guidance for formulating place-based and time-sensitive strategies to enhance station area vibrancy.
Vehicle re-identification (Re-ID) plays a vital role in vehicle-infrastructure cooperation systems, enabling continuous trajectory association across non-overlapping roadside camera views and providing essential perception redundancy for autonomous driving. However, in complex urban environments, frequent occlusions caused by surrounding vehicles, roadside structures, and pedestrians severely degrade Re-ID performance. Existing external model–based solutions often depend on auxiliary segmentation or detection networks, which increase computational overhead and hinder deployment in resource-constrained roadside units. In contrast, model-free approaches based on image enhancement provide limited gains when addressing structural occlusions, while multi-image feature fusion methods require additional images for joint training, substantially increasing training complexity. To address these challenges, this study proposes a novel dual-branch feature reconstruction model for the Re-ID (DFRR) framework, which integrates a dual-branch image feature encoder for discriminative feature extraction and an image reconstruction module for occluded image restoration under diverse and enhanced occlusion scenarios. Without relying on external models or neighboring image information, DFRR effectively recovers structural details in occluded regions, enabling robust single-image Re-ID in real-world traffic conditions. Extensive experiments on the VeRi-776 and DAIR-V2XReid datasets demonstrate the effectiveness of the proposed framework, achieving a higher mean average precision and delivering competitive performance compared with benchmark methods, particularly under severe occlusion conditions.
Efficient thermal management is paramount for enhancing the power density and operational longevity of proton exchange membrane fuel cells. This study systematically investigates twelve feasible coupled coolant–reactant flow-direction configurations within the same three-stage segmented cooling-channel geometry featuring straight, wavy, and tapered sections. We propose a robust data-driven framework that integrates high-fidelity physical modeling, Bayesian surrogate modeling, and multi-objective genetic algorithms to decouple complex thermo-fluid mechanisms. Case 12 is identified as the best-performing flow-direction topology, reducing the average temperature uniformity index by 45.7% relative to Case 8 based on ten Latin-hypercube-sampled coolant velocity–temperature difference combinations under identical geometric and operating conditions. The subsequent optimization determines a synergistic operating window with a velocity of 1.989 m/s and a temperature difference of 6.109 K, yielding an additional 14.53% enhancement in thermal uniformity while sustaining a high current density. The thermo-hydraulic synergistic mechanism involves reshaping the thermal boundary layer via induced secondary Dean vortices and spatial acceleration while successfully avoiding diminishing returns. This work provides a systematic design framework for coordinated flow-orientation and operating-parameter optimization in proton exchange membrane fuel cell thermal management systems.
Multimodal pathological diagnosis is a critical paradigm for accurate clinical decision making. Despite substantial progress, existing weakly supervised multimodal pathological diagnostic approaches integrating whole-slide images(WSIs) and structured clinical data typically rely on predefined collaboration strategies that assume uniform modality reliability, failing to account for sample-wise modality-specific noise and quality variations. Under weak supervision, the lack of fine-grained instance-level annotations prevents reliable assessment of sample-wise modality reliability, causing unreliable modalities to be indiscriminately fused and amplified, leading to unstable representation learning and unreliable multimodal decision making. To address this challenge, we propose a confidence-based multimodal diagnostic framework that models sample-wise modality reliability via prediction confidence and incorporates it throughout representation learning and multimodal decision making. Specifically, a dual-path global-local attention mechanism suppresses attention noise and stabilizes confidence estimation in pathological image representations, while large language models(LLMs)-based semantic reasoning improves the quality and consistency of structured clinical data through clinically plausible missing value imputation. At the multimodal decision stage, prediction confidence is used as an indicator of modality reliability to guide adaptive fusion, preventing unreliable modalities from dominating predictions. Evaluated on CAMELYON16, TCGA-BRCA, and TCGA-NSCLC datasets, the proposed model demonstrates robust improvements. On the TCGA-NSCLC dataset, it achieves 93.36% Accuracy and 96.40% AUC, outperforming the best unimodal baseline by 4.86% and 2.11%. Furthermore, it consistently surpasses multimodal approaches with up to 1.42% accuracy improvements, proving its superior robustness under sample-wise modality noise. GitHub repository: https://github.com/KuaLe/Pathologic_Fusion.
