Efficient Inventory Management for Manufacturers using Closed-loop Logistics and Dynamic Buffer Management

以往，製造商都是直接根據客戶給的預測量來進行生產，然而客戶的預測量往往大於實際的客戶需求量而產生大量存貨，因此製造商根據實際的顧客需求量去推測最佳庫存量，進而以此數量和客戶溝通而得到彼此認為的理想庫存量。為了讓製造商得知最佳的庫存量以做有效率的庫存管理，本論文分為兩個主題：第一個主題是閉迴路運籌，探討在生產過程中的可變動性，其對於製造量與庫存量的影響，透過發放問卷至臺灣某公司的在職人員來取得相關統計量，藉由 K—W 檢定與三種不同觀點的事後檢定，來檢驗各個可變動性的重要性，讓製造商能夠依照重要性的排序，做有效率的生產與庫存管理；第二個主題是動態庫存管理，製造商依照實際顧客需求量來決定最佳的擺放庫存量，為了最佳化持有庫存量以及緩衝庫存量，本論文分別針對遞增需求、遞減需求以及波動需求的產品，當實際庫存量超出緩衝庫存的情形發生時，在不同的產品生命週期以及前置時間下，利用粒子群演算法找到不同對象的最佳調整值。In the past, manufacturers produce products according to the vendor’s forecast. However, the vendor’s forecast is frequently higher than the actual customer’s demand so it caused numerous inventories. Therefore, manufacturers come to plan according to the actual customer’s demand to obtain the optimal quantity of inventory and use this consequence to negotiate with vendors to get both satisfaction results. This study therefore focuses on two topics: first topic is closed-loop logistics that discusses the flexibilities among production that might affect the quantity of products and inventory. A questionnaire was distributed to people who work in a Taiwanese company to obtain relevant data and test the level of importance of all flexibilities by K—W test and three different post-hoc tests so that the manufacturer is able to do production management and inventory management efficiently. Second topic is dynamic buffer management that is the actual application of planning based on the actual customer’s demand. To optimize the quantity of inventory kept and the buffer inventory set in dynamic buffer management, this research focuses on three cases including the increasing demand product, the decreasing demand product, and the fluctuation demand product to adjust different objects under the different conditions of life cycles and lead-times by particle swarm optimization.致謝................................................................................................................................... i 摘要 .................................................................................................................................. ii Abstract............................................................................................................................ iii List of Tables .................................................................................................................. vii List of Figures................................................................................................................ viii Nomenclature................................................................................................................... ix Chpater 1 Introduction................................................................................................. 1 1.1 Motivation ....................................................................................................... 1 1.2 Literature Reviews........................................................................................... 3 1.3 Contributions ................................................................................................... 7 1.4 Thesis Organization......................................................................................... 9 Chapter 2 Background Knowledge.............................................................................. 11 2.1 Closed-loop Logistics ( CLL ) ....................................................................... 11 2.1.1 The Effect Factors............................................................................... 12 2.1.2 Kruskal—Wallis Test ( K—W test ) ................................................... 15 2.1.3 Post-hoc Test....................................................................................... 16 2.2 Dynamic Buffer Management ( DBM )......................................................... 18 2.2.1 Particle Swarm Optimization ( PSO )................................................. 19 Chapter 3 Methodologies........................................................................................... 24 3.1 Closed-loop Logistics ( CLL ) ....................................................................... 24 3.2 Dynamic Buffer Management ( DBM )......................................................... 25 3.2.1 Initialization of DBM......................................................................... 26 3.2.1.1 The First Step in Initialization of DBM................................... 27 3.2.1.2 The Second Step in Initialization of DBM.............................. 28 3.2.1.3 The Third Step in Initialization of DBM................................. 31 3.2.2 The Object be Adjusted in DBM........................................................ 32 3.2.2.1 Adjustment for the Quantity of Inventory Kept ...................... 34 3.2.2.2 Adjustment for the Quantity of Buffer Inventory Set.............. 36 3.2.2.3 Adjustment for the Both Simultaneously ................................ 38 Chapter 4 Experiment Results ................................................................................... 42 4.1 The Effect Factors in CLL............................................................................. 42 4.2 The Adjustment Facotrs in DBM................................................................... 48 4.2.1 The Increasing Demand Product ........................................................ 49 4.2.2 The Decreasing Demand Product....................................................... 57 4.2.3 The Fluctuation Demand Product....................................................... 64 vi Chapter 5 Conclustions and Future Works ................................................................ 72 References ...................................................................................................................... 7