[[alternative]]Using KL Algorithm for Multi-Channel Wireless Broadcast Partitioning

[[abstract]] 隨著行動科技時代的來臨，無線通訊網路與行動裝置技術正以飛快的速度前進與發展，使行動使用者隨時隨地都能方便地獲得想要的資訊。在無線網路的環境之下，以目前的技術來看，無線網路技術的頻寬仍比有線網路技術的頻寬小很多；因此，廣播是一項節省頻寬浪費的有效技術！ 在多頻道的無線廣播環境中，將行動使用者所要求的資料項目平均分配到數個頻道上面，這樣一來可以縮短伺服器的廣播週期。當行動使用者所要求的資料集合分配在不同的頻道上，且位於相同的廣播時間點就會發生資料衝突。因為，發生資料衝突時，行動使用者必須等到下一個廣播週期才能再接收所要的資料項目，如此一來便會增加行動使用者資料項目的總存取時間。 在本文中，我們將KL演算法[18]運用於多頻道的無線廣播環境中，平均分配用戶端所請求的多重資料於多頻道上，以降低資料衝突發生的機率。從實驗結果顯示我們所使用的KL演算法可有效地降低資料發生衝突的機率：均勻分配比隨機分割法提升27%的效能、常態分配提升77%的效能，指數分配平均提升82%的效能。[[abstract]] With the coming time of mobile technologies, the Wireless Communication Network and the Mobile Unit Techniques are going fast step and amazingly development which make mobile technique user can get the desired information anywhere and anytime conveniently. Under the environment of wireless network, taking an overview of nowadays techniques, the bandwidth of wireless network techniques remain less than the wire network. Therefore, broadcast is a much more efficiency technique to save the use of bandwidth! In the environment of multi-channel wireless broadcast, configured the required data items on several channels averagely can make the interval of broadcast shorten. Whenever the data integration of users' desired been configured on different channel but on the same time-point of broadcast will the data conflict occur. Since the data conflict occurs, the mobile users have to wait for the next broadcast interval may them receive the desired data items. This will increase the total access time of data items of mobile users'. In this context, we take use of KL algorithm [18] expression at multi-channel wireless broadcast environment. Configuring the multiple data of client's required on multi-channel to decrease the rate of data conflict occurring. From the result of experience shows the KL algorithm expression what we take can effectively decrease the rate if data conflict: raising the performance of 27% at equalized configuration than random split, 77% at normalized configuration, and 82% at log configuration