| Summary: | 碩士 === 國立交通大學 === 工業工程與管理系所 === 101 === In the technical industries, cloud storage is a segment with highly attention for these days. Cloud storage is an application of cloud computing under infrastructure as a service (IaaS). Cloud storage is a virtualized pool of data center which operated by host companies and people can store data into it through internet. The progress of technology makes demands of cloud storage become higher. Unstable life cycles of products will also cause high variation of demands in cloud storage. The maintenance and purchasing cost of memory space need to be considered while avoiding the possibility of running out of it. Many researches were proposed to solve issues in cloud storage, but there is no research can improve the utilizations of memory space via inventory management. As a result, it is worthy to develop a memory replenishment strategy which can adapt to the environment of cloud storage. This research offers a MI hybrid memory replenishment strategy, which uses market information (MI) to build rolling forecast and then combines it with the demand pull and buffer management of theory of constraints (TOC). Both practical and simulative data are used to verify the effectiveness and feasibility of MI hybrid memory replenishment strategy. Furthermore, weighted moving average (WMA) and exponential smoothing (ES) are also used to build rolling forecast and then combine with TOC’s demand pull and buffer management (DPBM) separately. In the end, it is proved that MI hybrid memory replenishment strategy performs better than WMA hybrid memory replenishment strategy, ES hybrid memory replenishment strategy and TOC. MI hybrid memory replenishment strategy can decrease a large amount of on hand memory without impacting service level in the environment of cloud storage.
|
|---|
