近幾年來，在數值計算上講求的是計算效能，利用GPU提升計算速度也為目前重要發展之一，其中 CUDA 技術正是一個熱門的題材，倘若能善加利用此技術，那在數值計算上將會有很大的助益。因此，在本論文中，我們從中了解它的架構，並利用支援Matlab的JACKET軟件，對於數值求解非線性多變數最佳化問題，進行一些例子試驗，從中了解它的加速效果和數值結果，並與未加速的情況，進行比較，以及探討。

In recent years, a focus issue of scientific computing in the performance of the computation. A creative development for the high-performance computing is CUDA which adopts GPU to be an auxiliary device for the numerical computation, especially, parallel computation. We therefore survey the elementary architecture of CUDA in this thesis. A numerical experiment in GPU by using Matlab with software "Jacket" for solving nonlinear optimization problem is demonstrated. Some numerical results are illustrated the efficiency of GPU.

[1] 丁陳漢蓀, ” 嵌入式系統使用eForth” , 2003, 歐萊禮, pp.A6
[2] 張舒, 褚豔利, 趙開勇, 張鈺勃著, 陳乃塘譯, ”GPU 高效能運算之CUDA” , 2011, 碁峰,
[3] 賴炎連, 賀國平, ” 最優化方法” , 2008, 清華大學出版社, pp.220
[4] Han S. P., A Globally Convergent Method for Nonlinear Programming , J. Optimization Theory and Applications, Vol. 22, p. 297, 1977.
[5] Mokhtar S. Bazaraa, Hanif D. Sherali, C. M. Shetty, Nonlinear Programming: Theory And Algorithms , Third Edition, 2006, Wiley
[6] J. Nocedal and S. J. Wright, Numerical optimization , Second Edition, 2006, Springer
[7] Pavlos S. Geogilakis, Spotlight on Modern Transformer Design , 2009, Springer
[8] Powell, M.J.D., A Fast Algorithm for Nonlinearly Constrained Optimization Calculations ,Numerical Analysis, G.A.Watson ed., Lecture Notes in Mathematics, Springer Verlag, Vol. 630, 1978
[9] Ronald H.W. Hoppe, ”Optimization I”
[10] Wayne L. Winston, Munirpallam Venkataramanan, Introduction to Mathematical Programming: Applications and Algorithms , Fourth Edition, 2003, Cengage Learning
[11] Uri M. Ascher, Numerical Optimization , 2009
[12] ”http://developer.nvidia.com/category/zone/cuda-zone”
[13] ”http://www.accelereyes.com/”
[14] David Kirk, Wen-mei W. Hwu, ”http://courses.engr.illinois.edu/ece498/al/Syllabus. html” , 2007-2009
[15] ”http://www.smooth.url.tw/wiki/doku.php?id=cuda 教學”
[16] ”MATLAB-fmincon SQP Algorithm”