A Data-Driven Mixed Integer Nonlinear Programming Model for Cost-Optimal Scheduling of Perishable Production and Workforce

Abstract

This study presents a data-driven, Mixed Integer Nonlinear Programming (MINLP) framework for optimizing the multi-period production scheduling of perishable products with integrated workforce planning. Its primary novelty is the holistic integration of a continuous exponential decay function for product deterioration with dynamic workforce planning, creating a unified model that optimizes production, inventory, and labor simultaneously. This approach addresses key challenges in perishable inventory systems by treating labor as a controllable resource rather than a fixed constraint. Mathematically, the model includes nonlinear inventory balance equations with decay terms and resource-dependent capacity constraints. The objective is to minimize total operational cost, comprising production, holding, and spoilage costs. Computational experiments, based on a realistic case study, demonstrate that the proposed model reduces total system cost by 6.2% and spoilage costs by 43.2% compared to a standard heuristic benchmark. The resulting production and labor schedules align closely with demand fluctuations, supporting both economic and operational efficiency. This unified framework advances the mathematical modeling of sustainable production planning and offers a practical tool for real-world industries such as food processing and pharmaceuticals.