由零件供應商、製造商、配銷商、零售商以及客戶所組成的複雜供應鏈網路,及其所形成大量的資訊流會影響決策效率,甚至無法避免資料偏誤(Error-prone Data)或資料不完整(Incomplete Data);而其影響深遠,會造成前置時間增加、存貨水準不正確等一連串的連鎖反應,甚或形成「長鞭效應」。 本研究嘗試在交易網路 (Trans-Nets) 中之交易系統 (Trans System),提供計算系統估算之演算法及將演算法建置於 Trans-Nets中,作為供應鏈估測之工具;該工具係以動態估測之「卡爾曼濾波器」 (Kalman filter) 為主體,資料前、後處理以灰色系統理論(Grey system theory)中的累加生成 (Accumulated Generating) 序列及累減生成 (Inverse Accumulated Generating) 序列,來揭示資料潛在之規律性,依卡爾曼濾波器演算法及自迴歸法 (Auto Regression) 之結合,建構一具備處理資料不完整能力及提升資料精確性之即時動態修正的估測模式。
Abstract The supply chain and its considerable quantity information flow would affect decision efficiencies, but couldn’t avoid error-prone or incomplete data. The complex supply chain is composed of original suppliers, manufacturers, distributors, retailers, and customers. Incomplete information would result in mistaken chain reaction, which is the lead- time increasing and incorrect inventory level, even “bullwhip” effect. To provide the estimation function in supply chain operation, this study tries to provide an algorithm of mathematical calculation system in Trans-system of Trans-net. The core of the algorithm is Kalman filter. Pre- and post-processing of data reveal their pattern of data by accumulated-generating and inverse-accumulated- generating of Grey system theory. According to the combination of Kalman filter with auto-regression, we could build a dynamic estimation model with abilities of processing incomplete- data and promoting precision.