| Fabric of a particular color is typically purchased by apparel manufacturers in rolls of piece goods, each with the same commercial color designation. For these goods to be cut and sewn together requires a much tighter color tolerance than that specified by the commercial color designation. This is achieved by sorting the rolls into smaller groups. Shade sorting is defined as the identification of groups of like-colored materials which can be used together to make commercially acceptable products. Since fabric dyeing cannot be controlled completely, additional shade sorting of commercially acceptable goods is necessary. Generally, it is desired to sort the rolls into as small a number of groups as possible, with each group possessing no more than a prespecified maximum color range.; This research investigates numerical shade sorting systems and designs alternative methods which address the deficiencies of current systems. The shade sorting problem lies within a group of mathematically defined problems which are called NP-complete. No efficient solution procedures are known to exist. As an alternative to exact solution procedures, three heuristics are examined: Clemson Color Clustering (CCC), a graph labeling approach, and a set covering approach. These are compared by sorting several textile inventories and evaluating the shade groups produced.; A system of color inventory management based on the CCC method is developed. This system accounts for intra-roll as well as inter-roll color variation in the sorting procedure and provides a methodology for cutting operations. This system has been implemented in an industrial setting in cooperation with Tanner Companies, Inc. The success of the system is evident by improvements in color consistency, accounting, fabric utilization, and operational efficiency. |
