In Part I of the series, we describe stochastic computer models that simulate operations in the spinning, knitting, weaving, dyeing and finishing, and cut/sew sectors of the textile industry. The models are scaled to represent a supply chain designed to feed a garment-manufacturing operation involving four or five plants, i.e. part of each plant's output is ‘dedicated’ while simultaneously providing yarns and fabrics to the industry at large. Each of the sector models is unique because of the very different types of processing technology employed. The models are linked by means of streams of fabric orders from the manufacturing plants that make a range of garment types requiring many different fabrics for Basic (year-round sales), Seasonal (two or three seasons per year), and Fashion (shelf lives of 8–12 weeks) goods in a broad range of colors. In addition to each plant's product ranges and order sizes and frequencies, particular attention is paid to the machine-scheduling algorithms, although the models are deliberately kept at a ‘high’ as opposed to a ‘shop-floor’ level. The purpose of this modeling is to allow senior management to answer broad questions about the plants' ability to operate in a Quick Response environment. The various model outputs reflect this, having a heavy emphasis on on-time shipments, back-order levels, and service levels. In Part II of the series, we shall present the QR-related operating results to date, a description of a master-scheduling procedure to orchestrate the operations of the supply chain, ideas on an improved scheduling method, and an account of the construction of neural-network decision surface models as a decision support tool. We also overview ongoing efforts in technology transfer and in using ‘fuzzy’ mathematics to model the vagueness and uncertainty inherent in the supply- chain decision-making environment. The research effort of which this is a part is ongoing. We present these results in the hope of encouraging others to help carry the investigations forward.
Read full abstract