多維度偏好決策法應用於位置相依性資料預取查詢

隨著無線通訊技術之進步與行動裝置之蓬勃發展，使得行動計算環境逐漸成熟並衍生出位置相依資訊服務。其服務是指用戶於行動環境中具有可攜性之原則下，能在任何時間任何地點對遠端伺服器發出位置相依資訊查詢，而系統必須依據用戶所在位置之不同，而給於相對位置之查詢結果。在行動環境中，當行動用戶提出與當下位置相依物件查詢請求時，系統大都是以該次發出查詢的位置為絕對位置，計算出距離絕對位置最鄰近的查詢物件。如果系統單純計算其查詢物件之距離而並未考慮用戶偏好，會導致該系統回覆其查詢結果可能不符合用戶之期望。為此，本研究提出以多維度偏好決策法(Multi-dimensional Preference Decision-making Mechanism, MPDM)於計算行動用戶請求之偏好查詢處理，有別以往系統只能單純考慮查詢物件與用戶之空間屬性，該決策法還能考慮其查詢物件之非空間屬性是否符合用戶之偏好，達到用戶接收其查詢結果是滿足用戶查詢期望。另外亦利用代理伺服器之架構，依據用戶關注程度和偏好程度先行計算範圍內較可能查詢到之物件，預先給於行動用戶作資料預取之動作，以達到降低查詢回應時間與資料庫工作負載。效能分析上與探照燈式、路徑式、位置分割式資料預取進行效能評估，並定義其預取快取命中率、平均查詢回應時間、資料庫伺服器工作負載、快取資料使用率作為效能分析之評量指標。由模擬實驗結果得知在資料預取數量為5個且資料分佈數量為3個時，MPDM之預取快取命中率優於其他策略17%至40%。以查詢回應時間而言，MPDM比其他策略減少0.8秒至1.5秒。而MPDM於資料庫伺服器工作負載度之評估中比其它策略減少約44%至49%，在快取資料使用率方面，各策略於模擬環境中皆有不錯之效能。

關鍵字

偏好查詢；資料預取；位置相依資訊服務；代理伺服器

並列摘要

With wireless communication technology and the rapid development of mobile devices, making mobile computing environment to mature and derived location dependent information services. The service is a user in the mobile environment has the principle of portability, send location dependent information query to the remote server at anytime and anywhere, the system must be dependent on the location of the user, and reply to the query results of relative position. In a mobile environment, when mobile user presented with the current location dependent object queries, the system mostly based on the location of the queries where issued for the absolute position to calculated from the nearest neighbor query object. If the system is simple to calculate the query object but does not consider the preferences of user, will lead to reply to the query results may not meet user expectations. Therefore, this study proposes a Multi-dimensional Preference Decision-making Mechanism (MPDM) for processing the request about user preference. Not only the system consider the query of spatial attributes about objects and user’s location, but also the decision-making can consider non-spatial attributes about object with the user preference. So, that result that the user receive the query can match the user’s expectations. Another use the structure of proxy server, based on non-spatial attributes about user preference and the degree of attention to calculation what the object is more likely to inquired. It make user prefetch data ahead of time, so as to reduce query response time and minimize database workloads. In performance analysis, we compare MPDM with the headlight-based Data Prefetch (HDP), Partition-based Data Prefetch (PART) and Path-based Data Prefetch (PDP), and define four metrics: Prefetch cache hit rate (Phit), average query response time (Tar), database server workloads (Ldb), cache data used rate (Rdu) as the performance measuring. From simulation results, when the number of data prefetch is 5 and the number of data distribution is 3, MPDM was better than other strategies 17-40% on the Phit, In terms of Tar, MPDM was reduction 0.8 to 1.5 seconds than other strategies. MPDM reduction than other strategies about 44-49% on the Ldb. In respect of Rdu, the simulation results show any strategy has not bad performance in the simulation environment.

並列關鍵字

Preference query ； Data Prefetching ； Location dependent information services ； Proxy server

參考文獻

[50] 林彥亨, “以資訊關聯度為導向之近鄰預取查詢策略,” 國立台北科技大學碩士論文, July, 2009.

[1] J. E. Padgett, C. G. Gunther, and T. Hattori,“Overview of wireless personal communications,” IEEE Commun. Mag., vol. 33, pp. 28-41, Jan. 1995.

[2] D. L. Lee, J. Xu, and B. Zheng, “Data Management in Location-Dependent Information Services,” IEEE Pervasive Computing, vol. 1, Issue 3, pp. 65-72, 2002.

[4] M. H. Dunham and V. Kumar, “Location dependent data and its management in Mobile Databases,” in Proc. of the IEEE DEXA, pp. 414-419, Aug. 1998.

[5] E. Pitoura and G. Samaras, “Locating objects in mobile computing,” IEEE Knowledge and Data Engineering, pp. 571-592, Jul./Aug. 2002.

國際替代計量

多維度偏好決策法應用於位置相依性資料預取查詢

全文下載

主題瀏覽