Cutting optimization with variable-sized stock and inventory status data

Abstract

Many production environments require economical cutting of one-dimensional items according to bills of materials from objects of several standard lengths. However, even with optimized cutting substantial trim loss may occur. This trim loss should not be regarded as waste. It is returned to store and can be reused in future optimizations. Optimization of packing linear items into standard lengths is presented for items that cannot be packed into available lengths from inventory status data. The core of the proposed optimization tackles the variablesized bin packing problem (VBPP). The article presents a hybrid genetic algorithm that packs items into both available objects from the inventory and variablesized objects from the stock. The algorithm tries to minimize waste. Large trimloss items are returned as remnants to the inventory for subsequent optimizations.