English  |  正體中文  |  简体中文  |  Post-Print筆數 : 27 |  Items with full text/Total items : 94020/124474 (76%)
Visitors : 29062774      Online Users : 160
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://nccur.lib.nccu.edu.tw/handle/140.119/53421

    Title: 最佳化控制的加強及新的避鎖死計畫
    Other Titles: Enhancement of Optimal Control and a New Deadlock Avoidance Scheme
    Authors: 趙玉
    Contributors: 國立政治大學資訊管理學系
    Keywords: 派翠網;信標可控性;FMS;鎖死
    Petri nets;siphons;controllability;FMS;S3PR
    Date: 2010
    Issue Date: 2012-08-30 15:48:50 (UTC+8)
    Abstract: 摘要—向來,監督器是靠添加控制器到有問題信標來避免鎖死,會使很多系統性能降低。這將 導致狀態損失使得系統無法達到最大的允許。我們建議一個新的方案,從空虹吸狀態的系統恢 復前的活狀態。它因此可以到達原始未控制模型 (基於監督器之前從未達到) 相同的狀態數借 由類似于傳統的預防方法加入控制器 (和控制弧線)。與傳統的預防的方法不同,不生成新信 標。不過,添加了監視器的數目與有問題信標呈同樣指數網的大小的增長。此外,這方法由於 在空虹吸狀態後中止部分成品,會遭受物質損失。我們建議,著色一些控制弧形,以免物質損 失的一種無損方法。若要儘量減少控制器的數目,我們建議應用我們較早前的結果達到最優化 控制器的數目 (以及狀態數) 。就是按照正常序列的基礎,複合、 控制、 部分和完整的混合 物信標來添加控制器。我們又建議延伸使用每個控制器記錄虹吸管互補組中的標記數這一事 實,來逃避鎖死這種方法。在執行啟用的過渡t之前,我們將檢查t 輸出地點的一些虹吸管的控 制器 V的標記。如果 M(V) = M0(S)-1,觸發 t 會使S 無標記,則不允許t觸發。這比複雜的矩 陣計算或diagraphs更新為優。我們發現,傳統的一步看未來可能不足以避免所有鎖死和除了傳 統第二級別鎖 (SLD) ,還存在n 次級別 (n > 2) 鎖死。我們建議為 1) 執行進一步研究最 佳化控制借由延伸當前理論至 n > 3 強與弱依賴信標。和 2)研究新的避鎖死計畫。
    Abstract—Traditionally, monitors are added to problematic siphons to avoid deadlocks which degrade much system performance. This causes state losses making the system non-maximally-permissive. We propose a new approach which recovers the system from empty-siphon states to former live states. It therefore can reach the same number of states as the original uncontrolled model (never achieved before based on monitors) by adding monitors (and control arcs) similar to the traditional prevention approach. Unlike traditional prevention methods, no new siphons are generated. However, the number of added monitors grows exponentially with the size of the net since the number of problematic siphons does so. Also, the approach suffers from material loss by aborting partly finished products upon empty-siphon states. We propose a lossless approach to avoid material loss by coloring some control arcs. To minimize the number of monitors, we propose to apply our earlier results that the number of monitors (good states as well) can be optimized if one adds monitors in the normal sequence of basic, compound, control, partial and full mixture siphons. We further propose to extend this approach to deadlock avoidance using the fact that each monitor records the number of tokens in the complementary set of a siphon. Before firing en enabled transition t, we check the marking of monitors V at output places of t for some siphon S. If M(V)=M0(S)-1, firing t will make S unmarked, t is not allowed to fire. This has the advantage of neither complicated matrix computations nor updating of diagraphs. We discover that traditional one-step look ahead may not be enough to avoid all deadlocks and n-th level (n>2) deadlocks may exist in addition to the traditional second level deadlocks (SLD). We propose to 1) perform further research on our optimal control by extending the current theory to n>3 strongly and weakly dependent siphons. and 2) develop a novel deadlock avoidance scheme.
    Relation: 基礎研究
    研究期間:9908~ 10007
    Data Type: report
    Appears in Collections:[資訊管理學系] 國科會研究計畫

    Files in This Item:

    File SizeFormat
    992221E003.pdf653KbAdobe PDF476View/Open

    All items in 政大典藏 are protected by copyright, with all rights reserved.

    社群 sharing

    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback