請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/58324
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 王偉仲(Wei-Chung Wang) | |
dc.contributor.author | Jia-Hong Chen | en |
dc.contributor.author | 陳嘉宏 | zh_TW |
dc.date.accessioned | 2021-06-16T08:11:24Z | - |
dc.date.available | 2014-02-26 | |
dc.date.copyright | 2014-02-26 | |
dc.date.issued | 2014 | |
dc.date.submitted | 2014-02-19 | |
dc.identifier.citation | [1] Prasanna Balaprakash, Stefan M. Wild, and Boyana Norris. Spapt: Search problems in automatic performance tuning. Procedia Computer Science, 9(0):1959 – 1968, 2012. Proceedings of the International Conference on Com- putational Science, {ICCS} 2012.
[2] D. Bailey, R. Lucas, and S. Williams. Performance tuning of scientific applica- tions. Chapman & Hall/CRC Computational Science, 2010. [3] Gang Han, Thomas J. Santner, William I. Notz, and Donald L. Bartel. Pre- diction for computer experiments having quantitative and qualitative input variables. Technometrics, 51(3):278–288, 2009. [4] Curtis B. Storlie, Brian J. Reich, Jon C. Helton, Laura P. Swiler, and Cedric J. Sallaberry. Analysis of computationally demanding models with continuous and categorical inputs. Reliability Engineering & System Safety, 113(0):30 – 41, 2013. [5] Lophaven S. N., Nielsen H. B., and Sondergaard J. Matlab kriging toolbox dace. Technical Report IMM-TR-2002-12, Technical University of Denmark, 2002. [6] Lophaven S. N., Nielsen H. B., and Sondergaard J. Aspects of the matlab toolbox dace,. Technical Report IMM-REP-2002-13, Technical University of Denmark, 2002. [7] Peter Z. G Qian, Huaiqing Wu, and C. F. Jeff Wu. Gaussian process models for computer experiments with qualitative and quantitative factors. Technometrics, 50(3):383–396, 2008. [8] Qiang Zhou, Peter Z. G. Qian, and Shiyu Zhou. A simple approach to emulation for computer models with qualitative and quantitative factors. Technometrics, 53(3):266–273, 2011. [9] Donald R. Jones, Matthias Schonlau, and William J. Welch. Efficient global optimization of expensive black-box functions. J. of Global Optimization, 13(4):455–492, December 1998. [10] Zhiguang Qian, Carolyn Conner Seepersad, V Roshan Joseph, Janet K Allen, and CF Jeff Wu. Building surrogate models based on detailed and approximate simulations. Journal of Mechanical Design, 128:668, 2006. [11] Andy Keane. Surrogate-Based Optimization. John Wiley & Sons, Ltd, 2010. [12] Zhong-Hua Han and Ke-Shi Zhang. Surrogate-based optimization. [13] Alexander Forrester, Andras Sobester, and Andy Keane. Engineering design via surrogate modelling: a practical guide. Wiley, July 2008. [14] T. J. Santner, Williams B., and Notz W. The Design and Analysis of Computer Experiments. Springer-Verlag, 2003. [15] Ray-Bing Chen, Dai-Ni Hsieh, Ying Hung, and Weichung Wang. Optimiz- ing latin hypercube designs by particle swarm. Statistics and Computing, 23(5):663–676, 2013. [16] A. Grosso, A.R.M.J.U. Jamali, and M. Locatelli. Finding maximin latin hy- percube designs by iterated local search heuristics. European Journal of Oper- ational Research, 197(2):541 – 547, 2009. [17] M. Liefvendahl and R. Stocki. A study on algorithms for optimization of latin hypercubes. Journal of Statistical Planning and Inference, 136(9):3231 – 3247, 2006. [18] Edwin R. van Dam, Gijs Rennen, and Bart Husslage. Bounds for maximin latin hypercube designs. Operations Research, 57(3):595–608, 2009. [19] Peter Z. G. Qian and C. F. Jeff Wu. Sliced space-filling designs. Biometrika, 96(4):945–956, 2009. [20] Hengzhen Huang, Jian-Feng Yang, and Min-Qian Liu. Construction of sliced (nearly) orthogonal latin hypercube designs. Journal of Complexity, (0):–, 2013. [21] Jian-Feng Yang, C.Devon Lin, Peter Z.G. Qian, and Dennis K.J. Lin. Construc- tion of sliced orthogonal Latin hypercube designs. Stat. Sin., 23(3):1117–1130, 2013. [22] Jerome Sacks, William J. Welch, Toby J. Mitchell, and Henry P. Wynn. Design and Analysis of Computer Experiments. Statistical Science, 4(4):409–423, 1989. [23] V. Roshan Joseph and James D. Delaney. Functionally induced priors for the analysis of experiments. Technometrics, pages 1–11, 2007. [24] Nancy J. McMillan, Jerome Sacks, William J. Welch, and Feng Gao. Analy- sis of protein activity data by gaussian stochastic process models. Journal of Biopharmaceutical Statistics, 9(1):145–160, 1999. PMID: 10091915. [25] Lloyd N Trefethen and David Bau III. Numerical linear algebra. Number 50. Siam, 1997. [26] Javier R Movellan. Tutorial on gabor filters. Open Source Document, 2002. [27] P. Vanˇek, J. Mandel, and M. Brezina. Algebraic multigrid by smoothed aggre- gation for second and fourth order elliptic problems. Computing, 56(3):179–196, 1996. [28] William L. Briggs, Van Emden Henson, and Steve F. McCormick. A Multigrid Tutorial: Second Edition. Society for Industrial and Applied Mathematics, Philadelphia, PA, USA, 2000. | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/58324 | - |
dc.description.abstract | 自動化參數調校 (auto-tuning) 的問題在現今的科學計算領域是一個重要 議題。具有龐大運算量的科學計算方法倚靠著先進的電腦計算與模擬, 但由於當前的電腦架構越趨複雜,規模也越來越大, 其中許多的參數須經過調校後,方可獲得最佳計算效率。然而眾多的參數並不容易以手動或自動化的方式進行調校,尤其如果調校參數包含量化與類別變數 (qualitative and quantitative factors),問題的困難度更高。本研究針對包含量化與類別複合 型態的參數調校問題,提出不同種類的統計代理者模型 (surrogates)、以及不同的循序調校流程,達到自動化參數調校的目的。同時也藉由理論分析與數值實驗,比較分析各個方法的優缺點。再透過這些結果,提出一套模型選擇的方法與準則。 | zh_TW |
dc.description.abstract | The automatic performance tuning (auto-tuning) problem emerges in recent scientific computing applications. Usually, most of the applications are computationally intensive so that they rely on the computational power of the advanced computer. To achieve better performance, the performance tuning on related factors plays an important role. However, the architecture of modern computer becomes more and more complicated, so that the automatic performance tuning is indispensable. Meanwhile, the related factors involve various types, e.g. quantitative and qualitative factors. The difficulty here is the mixed types of input factors. We studied several statistical approaches (e.g. Gaussian Process model) to deal with such problems. A framework called surrogate-based tuning procedure is proposed, where the surrogate here means a statistical model of the tuning target. Moreover, our tuning procedure is an consecutive procedure, so an effective consecutive tuning procedure is necessary in this framework. To deal with the mixed input types, we proposed a extended method from the classical expected improvement method which is widely used in global optimization problems. And we compare their performances with many testing examples and real data in scientific computing. Finally, based on our results, we concluded a guideline for model selection in the auto-tuning procedure. | en |
dc.description.provenance | Made available in DSpace on 2021-06-16T08:11:24Z (GMT). No. of bitstreams: 1 ntu-103-R00221034-1.pdf: 10108563 bytes, checksum: a84b71f3cf5590fb37486c4e0f5c00a1 (MD5) Previous issue date: 2014 | en |
dc.description.tableofcontents | Abstract (in Chinese) i
Abstract (in English) ii Contents iv List of Algorithms v List of Figures vi List of Tables vii 1 Introduction 1 2 Methods and Algorithm 3 2.1 Surrogate-basedAuto-tuningAlgorithm . . . . . . . . . . . . . . . . 3 2.2 InitialDesign ............................... 5 2.3 SurrogateConstruction.......................... 6 2.3.1 IndependentGaussianProcess.................. 6 2.3.2 Q&QGaussianProcess..................... 9 2.3.3 Comparison between IGP and QQGP ............. 13 2.4 OptimaInvestigation........................... 15 2.4.1 Minimum-shared Expected Improvement (MSEI) . . . . . . . 16 2.4.2 MinimalPrediction(MP) .................... 19 3 Numerical Results 21 3.1 Examples and Experiment settings ................... 21 3.1.1 Examples ............................. 21 3.1.2 Experiment settings ....................... 29 3.2 Performance Comparison and Results.................. 30 3.2.1 Robustness ............................ 31 3.2.2 Efficiency ............................. 36 4 Discussion 50 4.1 Robustness Comparison ......................... 50 4.2 EfficiencyComparison .......................... 52 4.3 Conclusion of Performance Comparison ................. 53 5 Conclusion 54 6 Acknowledgement 56 References 59 Appendix A Proofs 60 A.1 Part I: Proof of Expected Improvement properties . . . . . . . . . . . 60 Appendix B True surfaces of testing problems 63 | |
dc.language.iso | en | |
dc.title | 透過統計的代理模型對包含量化與類別參數的問題進行自動化參數調校 | zh_TW |
dc.title | The Auto-tuning Procedure to the Problem with Quantitative and Qualitative Variables via a Statistical Surrogate-based Model | en |
dc.type | Thesis | |
dc.date.schoolyear | 102-1 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 陳瑞彬,陳瑞彬 | |
dc.subject.keyword | 量化與類別變數,電腦實驗,統計代理者模型,自動化參數調校, | zh_TW |
dc.subject.keyword | Qualitative and quantitative factors,Computer Experiment,Gaussian Process,Infill criterion,Expected Improvement,Surrogate modelling,Performance Tuning, | en |
dc.relation.page | 72 | |
dc.rights.note | 有償授權 | |
dc.date.accepted | 2014-02-19 | |
dc.contributor.author-college | 理學院 | zh_TW |
dc.contributor.author-dept | 數學研究所 | zh_TW |
顯示於系所單位: | 數學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-103-1.pdf 目前未授權公開取用 | 9.87 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。