In skill-constrained production-inventory systems, the available qualified human capacity tomorrow hinges on today's training decisions: production requires certified workers, certifications decay unless maintained, and training consumes the same scarce worker hours needed for production now. We study a closed-loop skill-constrained model predictive controller that, at every shift, solves a finite-horizon mixed-integer program over production, inventory, backlog, and training, featuring binary predicted certification, hard production eligibility, and an interpretable terminal value that prices certified-capacity gaps at the horizon boundary; only the first-period action is executed before replanning.
On synthetic, seed-controlled SkillChain-Gym scenarios—encompassing announced and surprise new-skill shocks, demand shocks, absenteeism, forecast- and availability-quality modes, capacity-boundary and training-rate sweeps, and negative controls—we evaluate the controller against production-only and maintenance-only ablations, static cross-training insurance plans, and a robust reactive heuristic, all under an ex-ante locked configuration and paired statistics. The results indicate regime dependence rather than superiority: no policy class dominates. Predictive control is beneficial when skill or labor bottlenecks are forecastable early enough for training to complete; however, lean static insurance remains hard to beat under surprise shocks, near the demand-capacity boundary, and wherever pre-shock slack makes insurance cheap. Attribution ablations distinguish between certification maintenance, re-acquisition of lapsed certifications, and greenfield skill acquisition. Ultimately, forecastability—not adaptivity per se—determines when predictive control is advantageous.
Blogger's Review: This study delves into the intricate relationship between skill constraints and production efficiency, highlighting the significance of strategic training and certification in enhancing supply chain resilience. The application of model predictive control allows for optimized resource allocation amid uncertainty, yet the response to unforeseen events requires further research and practical validation.