In recent years, the escalation in emergency occurrences has underscored the pressing need for expedient responses in delivering essential supplies. Efficient integration and precise allocation of emergency resources under joint government–enterprise stockpiling models are pivotal for enhancing emergency response effectiveness and minimizing economic repercussions. However, current research predominantly focuses on contract coordination and cost-sharing within these joint reserve modes, overlooking significant discrepancies in emergency supply classification standards between government and enterprise sectors, as well as the asymmetry in cross-sectoral and cross-regional supply information. This oversight critically impedes the timeliness and accuracy of emergency supply responses. In practice, manual judgment has been used to match the same materials under differing classification standards between government and enterprise reserves. Still, this approach is inefficient and prone to high error rates. To mitigate these challenges, this study proposes a methodology leveraging the BERT pre-trained language model and TextCNN neural network to establish a robust mapping relationship between these classification criteria. The approach involves abstracting textual representations of both taxonomical classes, generating comparable sentence vectors via average pooling, and calculating cosine similarity scores to facilitate precise classification mapping. Illustrated with China’s Classification and Coding of Emergency Supplies standards and Global Product Classification standards, empirical validation on annotated data demonstrates the BERT-TextCNN model’s exceptional accuracy of 98.22%, surpassing other neural network methodologies such as BERT-CNN, BERT-RNN, BERT-BiLSTM, etc. This underscores the potential of advanced neural network techniques in enhancing emergency supply management across diverse sectors and regions.

An emergency collection plan may perform poorly because of its visit order, its supplier combination, or both. Separating these losses helps determine how much reordering alone can improve the plan. We study collection from distributed inventories under a prescribed policy: each pickup meets remaining demand subject to available stock, and visits stop once all material requirements are satisfied. Conditional optimization considers only orders that reproduce exactly the same realized supply-point set. It separates recoverable ordering loss from activation-set loss, the gap between that set's conditional optimum and the global optimum. We examine these losses using complete exact landscapes, parameter comparisons, and individual-plan diagnostics. Across ten complete landscapes with eight or ten candidate supply points, activation-set loss dominates under set-balanced aggregation, with median instance-level shares of approximately 79\% and 72\%, respectively. Optimal reordering therefore leaves most aggregate loss unresolved in these landscapes. Some instances and plans nevertheless retain substantial ordering opportunities. Across 70 exact parameter scenarios, emphasizing configuration-switching or loading/unloading costs increases the sequence-weighted ordering-loss share, while emphasizing coordination costs reduces it. Responses to demand changes vary across instances; activation-set loss remains dominant in the tested scenarios. We interpret these responses by examining admissible orders, realized supply allocations, and cost changes under set-preserving reordering. Individual plans can also be assessed when the global optimum is unknown, as illustrated by search-output diagnostics and a case application. Exact conditional optimization determines the maximum fixed-set improvement; heuristic search identifies attainable gains without certifying that maximum. In the case application, another feasible combination achieves an objective value below the retained set's exact conditional optimum. Matching that performance requires changing the combination. These findings suggest evaluating supplier combinations by their best admissible execution and reassessing each plan when demand or priorities change. Current objective values and aggregate loss proportions alone do not establish how much an individual plan can improve.