|
|
|
| Quality optimizing real-time scheduling for incremental context services |
| YANG Zhao-hui1, LI Shan-ping1, LIN Xin1,2 |
1. College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China; 2. Department
of Computer Science and Technology, East China Normal University, Shanghai 200241, China |
|
|
|
Abstract A two phase task model was proposed in order to achieve optimal user experience for an incremental context service under resource constraints and with context timeliness requirements. The model divides each job of an incremental service into a primary part that provide the initial service and a optional part that provide the improvements. A corresponding two level scheduling theme was designed to utilize the model. Several scheduling algorithms were designed to execute the primary parts in a timely manner and the optional parts in a value maximizing manner by analyzing the schedulability of the primary parts and the value generation pattern of the optional parts. Simulation results show that all the three primary part scheduling algorithms can meet context timeliness requirements under resource constraints. The performances of the optional part scheduling algorithm were compared when using different value prediction methods. Results show that the simpler delta value method can achieve performance close to that of the fit curve method, and is more suitable for practical use．
|
|
Published: 22 February 2012
|
|
|
增量型上下文信息服务的质量优化实时调度
为了优化增量型上下文服务的用户体验,同时兼顾服务资源约束和上下文时效性需求,提出将增量型上下文服务作业分为必须完成的基本段和提供增量的延伸段的两阶段任务模型,及与之相应的两级调度模型. 通过分析基本段的可调度性和延伸段价值（量化的用户体验）产出随时间变化的趋势,分别设计针对基本段和延伸段的多种调度算法,在保证基本段按时完成的同时,优化延伸段提供的总价值. 通过模拟实验验证了3种基本段调度算法均同时满足服务资源约束和上下文时效性.比较不同价值估算方法对延伸段调度算法性能的影响,发现相对简单的价值差估算法性能接近拟合曲线法,比后者更实用.
|
|
| ［1］ XU J, TANG X, HU H, et al. Privacyconscious locationbased queries in mobile environments ［J］. IEEE Transactions on Parallel and Distributed Systems, 2010, 21(3): 313-326．
［2］ DARGIE W. The role of probabilistic schemes in multisensor contextawareness ［C］∥5th Annual IEEE International Conference on Pervasive Computing and Communications Workshops.White Plains: IEEE, 2007: 27-32．
［3］ KWANGEUN K, KWEEBO S. Development of context aware system based on Bayesian network driven context reasoning method and ontology context modeling ［C］∥ International Conference on Control, Automation and Systems. Seoul: ICROS of Korea, 2008:2309-2313．
［4］ YE J, COYLE L, DOBSON S, et al. Using situation lattices to model and reason about context ［C］∥ 4th International Workshop on Modeling and Reasoning in Context.Roskilde: Roskilde University, 2007: 1-12．
［5］ HAN Q, VENKATASUBRAMANIAN N. Addressing timeliness/accuracy/cost tradeoffs in information collection for dynamic environments ［C］∥ 24th IEEE International RealTime Systems Symposium.Los Alamitos: IEEE, 2003: 108-117．
［6］ HAN Q, VENKATASUBRAMANIAN N. Timelinessaccuracy balanced collection of dynamic context data ［J］. IEEE Transactions on Parallel and Distributed Systems, 2007, 18(2): 158-171．
［7］ 林欣, 李善平, 杨朝晖. 新鲜度敏感的上下文推理实时调度算法 ［J］. 电子与信息学报, 2009, 31(5): 1185-1188．
LIN Xin, LI Shanping, YANG Zhaohui. Freshnessaware realtime scheduling algorithm for context reasoning ［J］. Journal of Electronics and Information, 2009, 31(5): 1185-1188．
［8］ LIU J W. Realtime systems ［M］. Upper Saddle River: Prentice Hall, 2000: 130-140．
［9］ Bayesian network tools in Java (BNJ). ［20101101］. http:∥bnj.sourceforge.net/.
［10］ CHENG J, DRUZDZEL M J. AISBN: an adaptive importance sampling algorithm for evidential reasoning in large Bayesian networks ［J］. Journal of Artificial Intelligence Research (JAIR), 2000, 13(1): 155-188． |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
