Glass container production planning with warm-ups and furnace extraction variation losses

Abstract

This article considers the challenging problem of production lot sizing and scheduling that arises in the glass container industry. In such manufacturing environment, a furnace operates continuously feeding its attendant container forming machines with melted glass. Setups in the aforementioned machines interrupt the melted glass flow and, hence, the glass paste undergoes a heat loss. Consequently, a warm-up process is needed after each setup changeover. In addition, setups can affect the furnace output (total amount of glass extracted from the furnace) resulting in extraction variation losses. Both warm-ups and furnace extraction variations prevent the capacity of machines from being fully used and incur in additional costs. This study addresses such specificities for the first time in lot sizing and scheduling literature by introducing a novel integer programming model, which represents the problem appropriately, and a variable neighborhood search (VNS) solution method. A relevant case study conducted in an amber glass bottle plant shows that the model is able to comprise the problem requirements, and the VNS produces good-quality solutions in reasonable computational times. The reproducibility of the results is ensured by additional experiments involving artificial instances.