Collaborative Research: Generative AI for Autonomous Composition of Resilient and Explainable Service-Oriented Supply Chains

Manufacturing supply chains are increasingly vulnerable to disruptions caused by natural disasters, geopolitical instability, demand fluctuations, and complex dependencies among globally distributed production networks. Improving the resilience and adaptability of these supply chains is essential for maintaining economic competitiveness, technological leadership, and national security. At the same time, many manufacturers, particularly small and medium-sized enterprises, face barriers to participating in digitally connected supply networks due to fragmented information, limited visibility into available capabilities, and the lack of scalable mechanisms for identifying and coordinating manufacturing partners. This research advances a service-oriented vision of manufacturing in which production capabilities can be discovered, combined, and reconfigured as interoperable services across a distributed ecosystem. By developing methods that enable the autonomous creation of resilient and transparent supply chain configurations, the project supports more agile responses to disruptions and changing market conditions. The resulting infrastructure will strengthen domestic manufacturing ecosystems, increase opportunities for collaboration among manufacturers, and improve access to advanced digital technologies. Educational activities will enhance manufacturing and artificial intelligence curricula through new instructional materials, microcredentials, research experiences, outreach programs, and community engagement activities designed to increase participation in advanced manufacturing and computing fields. The specific goal of this research is to develop a computational framework that integrates semantic technologies and generative artificial intelligence to enable the autonomous composition of resilient and explainable service-oriented supply chains. The research will establish a formal ontology for representing manufacturing services, their capabilities, constraints, and logical dependencies, creating a common foundation for interoperability across diverse manufacturing environments. Using this ontology, the project will develop methods for constructing large-scale manufacturing service knowledge graphs by transforming manufacturer capability information into machine-interpretable service representations. The research will also create generative artificial intelligence techniques that extract required manufacturing services from work orders and automatically compose feasible multi-tier supply chain configurations from dynamic networks of manufacturing service providers. Novel resilience assessment methods based on network perturbation analysis will be developed to evaluate the robustness and adaptability of generated supply chain configurations under disruption scenarios. In addition, the project will investigate approaches for improving the transparency and trustworthiness of artificial intelligence systems through evidence-based explanations, counterfactual analysis, and information-theoretic methods that identify the factors influencing service graph construction and supply chain recommendations. The resulting framework will establish foundational methods for scalable, interoperable, resilient, and explainable service-oriented manufacturing ecosystems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.