基于嵌套分割算法的復雜物流系統(tǒng)庫存仿真優(yōu)化.pdf

1、華中科技大學碩士學位論文基于嵌套分割算法的復雜物流系統(tǒng)庫存仿真優(yōu)化姓名：葉明基申請學位級別：碩士專業(yè)：系統(tǒng)工程指導教師：朱衛(wèi)鋒20090527華 中 科 技 大 學 碩 士 學 位 論 文 II Abstract Complex Logistics (CL) is consisted of raw material suppliers, manufactures, products suppliers and retailers. H

2、ow to configure the Order-up-to level of the entities under multi-period review inventory control strategy is the main problem in CL. Different order-up-to levels produce different inventory cost and fill rate. Inventory

3、 cost shows the performance of CL, and it always includes holding cost and shortage cost. The aim of the thesis is to find the balance of holding cost and shortage cost to make the CL system perform at a low total cost l

4、evel with an acceptable fill rate. As there are many uncertainties in CL, the traditional analytical method can not be used to get proper order-up-to level to meet the lowest cost in the supply chain. The simulation opti

5、mization is considered as the most usable and feasible method to solve this kind of random system with many complexities. A new centralized simulation architecture for CL combined nested partitions algorithm called CLCSi

6、m-NP is put forward based on the existing CL centralized simulation architecture-CLCSim. CLCSim-NP is an optimization architecture integrated with the nested partitions algorithm to solve the order-up-to level allocation

7、 in CL. The procedures for solving CL order-up-to level using CLCSim-NP are proposed. After analyzing the advantages and the solving procedures of CLCSim-NP, the simulation software for CLCSim-NP is designed and implemen

8、ted. A CL example with two manufactures, two suppliers and three retailers is taken by using CLCSim-NP simulation software to solve the order-up-to levels of the seven entities. The CL considered here adopts the (t, S)

9、inventory control strategy and has random demands. The total cost of the CL is collected after some sets of simulation experiments. The SPSS is used to do the data analysis by calculating the mean, the standard error of

10、mean, the standard deviation and the range. It shows that simulation runs and selecting region strategies have impacts on the solutions of order-up-to level. The relationship between reliability and simulation runs is al