A manufacturing capability-aware supplier decision framework for resilient production allocation and disruption response
Year: 2026
Type: article
Abstract: This study presents a manufacturing capability-aware supplier decision framework that connects supplier identification, production allocation, and disruption-responsive reconfiguration within a unified computational framework. Unlike conventional supplier selection and supplier discovery approaches that often assume feasible suppliers are already known or focus primarily on matching, ranking, or business-level criteria, the proposed framework represents supplier feasibility using integrated product geometry and manufacturing-related attributes. A multimodal autoencoder is used to learn product representations that reflect manufacturing capability, and feasible suppliers are identified through distance-based thresholding in the learned representation space. The identified supplier pool is then passed to downstream optimization for demand allocation under capacity constraints and reconfiguration under supplier-side disruption. Experimental results show that the proposed method improves supplier identification performance, achieving 0.99 precision, 0.98 recall, and 0.98 F1-score. This improvement leads to more executable downstream decisions, as reflected by lower effective objective ratios that penalize unmet demand and capability-infeasible allocations relative to ground-truth reference objective values. Under supplier capacity shortage and degradation disruptions, the proposed framework yields more reliable and executable demand reallocation outcomes under a common capacity-constrained allocation and reconfiguration model. These results show that manufacturing capability-aware supplier identification should be evaluated not only as a matching task, but also as an upstream decision stage that directly shapes downstream production allocation and disruption-responsive supplier reconfiguration decisions.
Source: Journal of Manufacturing Systems
Authors: Suyoung Park, Shreyes N. Melkote
Institution: Georgia Institute of Technology
Language: English
Subfield: Strategy and Management
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $3,850
Funder: National Science Foundation
OpenAlex ID: W7213655876
