烘焙代工業之排程研究

近年由於網紅經濟崛起，烘焙業代工的型態開始改變，廠商需要協助客戶研發更多品項，烘焙業代工已走向精緻烘焙代工，訂單少量多樣。廠商研發新產品的能力，且能大量生產客製化的產品。這樣形態下的生產排程也比機械化產線，大規模製造，少樣多量的生產排程複雜許多。個案公司為精緻烘焙代工產業中的代表性公司，雖然每年營收成長，但其少量多樣的產品型態，人力比重高(因產品手續繁複，產品種類迭代快，無法有全自動的生產線)的生產模式，在生產排程上遇到瓶頸，除了管理者花費大量時間精力在排程，採用小心謹慎的接單策略，避免訂單延遲。但也因此減少了潛在的營收。本研究以個案研究法和數學規劃模型，訪談公司經營者，實地調查產線，整理出精緻烘焙代工產業特性。回顧國外食品廠和乳製品廠的相關排程文獻，建構出烘焙產業的混合整合線性規劃模型，以實際案例測試結果，並於個案公司內部展示說明績效，其中，因為人工排程不能接的訂單，數學規劃模型仍可以排出可行的排程。除此之外，本研究也整理個案公司給予的回饋，敘述說明管理上的幫助。

關鍵字

烘焙業；生產排程；網紅經濟

並列摘要

In recent years, due to the rise of the Internet celebrity economy, the type of OEM in the baking industry has begun to change. Manufacturers need to assist customers in developing more products. The OEM in the baking industry has moved towards a refined model, with small and diverse orders. Manufacturers need to build the ability of to develop new products and mass-produce customized products. The production schedule in refined model is also much more complicated than the mechanized production line, large-scale manufacturing, and a small variety of production schedules. The case studied in this thesis is a representative company in the refined baking OEM industry. Although its revenue grows annually, its small and diverse product types require a high proportion of manpower as well. This is because of the complicated product procedures and rapid iteration of product types. A fully automatic production line in the production line is very costly for this company. The production schedule, however, faces some bottlenecks. Managers not only need to spend a lot of time and energy on scheduling, they also have to adopt a careful order-taking strategy to avoid order delays at the cost of reducing potential revenue. This research uses the case study method and MILP (Mixed Integer Linear Programming), interviews company operators, and conducts field investigations on production lines, to sort out the characteristics of the refined baking OEM industry. We also review the relevant scheduling literature of foreign food factories and dairy product factories, construct a MILP model for the refined OEM baking industry, test the results with actual cases, and demonstrate the performance in the case company. Our results show that the MILP model can produce a feasible schedule. We also summarize the feedback given by the case companies, and discuss the managerial implication.