應用基因演算法設計資料倉儲與效能分析

Research of Applying Genetic Alogrithm to Design Data Warehouse and Performance analysis

劉文義(Wen-Yi Liu)

指導教授 ： 陳凱瀛

國立臺北科技大學/管理學院/工業工程與管理系所/碩士(2005年)

https://doi.org/10.6841/NTUT.2005.00024

摘要

維修管理一直是整體製造上所重視一環，維護策略主要的目的就是要降低維護成本、協助維護人員在正確的時間和地點做出正確決定，並且讓所有的設備狀態都能夠保持在控制的範圍之中。而現行的電腦化維修管理系統(Computerized Maintenance Management System, CMMS)，亦可稱為設備管理系統，所產生的資料都為詳細的交易資料並不適合做為決策分析和各項整合應用。
因此，為了提供製造業設備快速且精確的整合資訊，本研究應用資料倉儲（Data Warehouse, DW）概念來整合設備管理相關資訊，針對半導體封裝產業為研究對象，依照使用者需求對設備管理中的設備維護、備品管理兩模組進行資料倉儲模型的建立和資料方體的設計。資料倉儲建立上，利用SQL 2000及IIS等相關工具將設備管理系統資料倉儲予以實現，讓使用者透過WEB的方式，依據所需的資訊在線上做多維度的查詢。透過本系統的建立，能夠提供業界或者是相關研究者，ㄧ個完整而且符合效能的設備管理系統資料倉儲，做為往後設計和建構的依據。
資料方體(Data Cube)的挑選上，採用部分實體化(partial materialized)方式來實建資料方體，利用貪婪法(Greedy Algorithm)和兩項基因演算(FGS+GA、BGS+GA)做為資料方體挑選依據，藉由考量查詢成本、維護成本等因素，挑選出最適當的資料方體予以實建。在資料方體挑選方法比較中，由實驗結果，發現由本研究所提出的FGS+GA在大部分的空間限制條件下，其實體化方體總成本與BGS+GA差異不大，兩者皆有優於傳統貪婪法(Greedy algorithm)的求解品質；但在求解效能上，FGS+GA相較於BGS+GA能大大的縮減處理時間。在資料倉儲系統效能分析上，針對本研究所建立資料方體在使用者查詢時間和空間使用率比較發現，由本研究所採用方法的挑選結果，空間使用率相對於完全實體空間可以降低70%，並且在查詢時間上能有所改善。

關鍵字

基因演算法 ； 資料倉儲 ； 線上分析處理 ； 資料方體 ； 設備管理系統

並列摘要

Today, maintenance management has been one of the most important tasks in manufacturing. The aims of maintenance strategies are to reduce maintenance costs while improving maintenance operation and to help the maintenance managers to make the right decision at the right time in the right place. Computerized maintenance management systems(CMMS) can help us to deal with these maintenance tasks, but doesn’t have Decision Support System (DSS) capability.
In this thesis, we design a data warehouse for Equipment Management System(EMS) to help the manager inquire and make decision by different dimension in a quickly and flexible way. Many cubes, such as MTTR, MTBF, Spare response time,… are created using the flake schemas. Finally, some concluding remarks are made. The results of this research will also provide a reference model for enterprise when EMS data warehouise is evaluated as a To Be system.
For data cube selection, in order to minimize query time under the storage limit, we adopt partial materialized way to select data cubes. In this thesis, we use three ways to select data cubes：Greedy algorithm, FGS+GA, which incorporate genetic algorithm with forward greedy algorithm and BGS+GA, which incorporates genetic algorithm with forward greedy algorithm. According to our experiments, in case of most of storage constraint, the solution generated by FGS+GA is as well as the solution generated by BGS+GA, and both of them is superior to that found by Greedy algorithm. Furthermore, when we compare the efficiency of FGS+GA with BGS+GA, we find that the efficiency of FGS+GA is superior to BGS+GA.
For performance anlysis, we compare the query time and space utilication of the selected cubes obtained by method we proposed with all materialized way. According to our experiments, the results obtained from our method can decrease 70% of storage space, and the query time of our method is as well as all materialized way.

並列關鍵字

Genetic Alogrithm ； Data Warehouse ； Equipment Management System ； On-Line Analytical Processing(OLAP) ； Data Cube

參考文獻

[35]沈育樹，宮大川，“整合性維護管理系統設計”,工業工程學刊,第20輯,第337－354頁，2003。

[3]C.I. Ezeife, “A uniform approach for selecting views and indexes in a data warehouse,” International Database Engineering and Applications Symposium, 1997, pp. 151-160.

[5]H. Jiawci, K. Micheline, “Data mining Concepts and Techniques”, Morgan Kaufmann, 2001, pp.71-79

[6]H.J. Watson, A.D. Annino, H. Barbara, et al. “Current Practices in Data Warehousing,” Imformation Systems Management, Vol.18 Issue 1, Winter2001, pp.47-55.

[7]H. Zhihong, Z. Zhixue, “CMMS and its application in power systems”, IEEE International Conference on Systems, Man and Cybernetics, Vol.5, 5-8, 2003, pp.07-12.