Abstract 

This study addresses the complex cutting stock problem encountered by steel-pipe manufacturers. Key features of this problem include raw material–product eligibility, multi-period constraints, and inventory level restrictions. The primary objective is to allocate product orders to eligible raw materials while optimizing the production plan over multiple time periods. To tackle this challenge, we propose a comprehensive mathematical model with an objective function that minimizes raw-material costs, product inventory levels, the use of over-specified raw materials, unfulfilled orders, and excess inventory while adhering to eligibility constraints and yield limitations. To solve the model, we introduce a large neighborhood search algorithm integrated with a heuristic initial solution construction, destroy operators, and repair operators. A series of experiments on randomly generated instances and real-world data demonstrate the algorithm's effectiveness, outperforming conventional mathematical model-based approaches. For large-scale real-world problems, our method achieved a 24.8 % reduction in cost and a 14.0 % improvement in inventory optimization compared to current company practices