TITLE:
Predictive Risk Modeling and Optimization for Resilient Semiconductor and Defense Manufacturing Supply Chains: A Systematic Review
AUTHORS:
David Ishimwe Ruberamitwe, Patrick Ishimwe
KEYWORDS:
Supply Chain Resilience, Risk Modeling, Optimization, Semiconductor Manufacturing, Defense Industry, Predictive Analytics, Disruption Management, PRISMA Systematic Review
JOURNAL NAME:
Journal of Computer and Communications,
Vol.14 No.4,
April
17,
2026
ABSTRACT: Background: Semiconductor and defense manufacturing supply chains face unprecedented disruptions from geopolitical tensions, natural disasters, and global pandemics. These critical industries require robust predictive risk modeling and optimization approaches to ensure supply chain resilience and operational continuity. Despite growing research interest, a comprehensive synthesis of methodologies, findings, and gaps in this domain remains lacking. Methods: We conducted a systematic review following PRISMA 2020 guidelines. Six electronic databases were searched (Dimension AI, Google Scholar, ArXiv, PubMed) from inception to February 2026, yielding 960 initial records. After deduplication (273 unique papers) and systematic screening with predefined inclusion/exclusion criteria, 30 studies were included in the full synthesis. Data extraction focused on study characteristics, risk modeling approaches, optimization methods, resilience strategies, key findings, and limitations. Quality assessment examined methodological rigor, data sources, and validation approaches. Results: The 30 included studies employed diverse methodologies, including stochastic programming (n = 8), machine learning and AI-driven approaches (n = 7), simulation-based methods (n = 6), multi-objective optimization (n = 5), and network-based models (n = 4). Risk modeling approaches predominantly addressed demand uncertainty (73%), supply disruptions (67%), yield variability (43%), and geopolitical risks (27%). Optimization techniques included genetic algorithms, mixed-integer programming, reinforcement learning, and multi-stage stochastic optimization. Key resilience strategies encompassed supplier diversification, inventory buffering, flexible capacity allocation, and real-time monitoring systems. Studies demonstrated significant improvements in supply chain performance metrics, with disruption cost reductions ranging from 15% to 45% and service level improvements of 10% to 35%. Conclusions: Current research demonstrates substantial progress in predictive risk modeling and optimization for semiconductor and defense supply chains, with stochastic programming and AI-driven approaches showing particular promise. However, significant gaps remain in integrated frameworks that combine multiple risk types, include real-world validation studies, and account for emerging threats such as cyber-physical attacks. Future research should prioritize industry-academia collaboration, development of standardized benchmarks, and incorporation of sustainability considerations alongside resilience objectives. These findings provide actionable insights for supply chain managers, policymakers, and researchers working to enhance the resilience of critical manufacturing supply chains.