IE 468: Optimization Methods and Models (Spring 2019)
优化方法代写 Project requirements: The project is an integral part of IE 468, focusing on the devel- opmemt of some optimization algorithms
Project (40%) Uday V. Shanbhag
Project requirements: The project is an integral part of IE 468, focusing on the devel- opmemt of some optimization algorithms in the context of continuous optimization. The course will comprise of providing a typed report (with m-files and iteration logs in the appendix) that contains the following.优化方法代写
Outline of report.
Section I: Unconstrained optimization.优化方法代写
(i)A description of each of the four algorithms: (a) Gradient methods with exact line search; (b) Gradient method with backtracking line search; (c) Diagonally scaled gradientmethod with exact/backtracking line search; (d) Hybrid Newton (Most of this is already done)
(ii)Thisdescription should contain a short description of the method and provide state- ments regarding both the convergence of the iterates and the associated rate of con- vergence (if any).优化方法代写
(iii)Finally,each algorithm should be applied to Exercise 3 (quadratic minimization) and Example 5.8 (Rosenbrock’s function) and the report should contain a log of iterations where f (xk) and xf (xk) are tabulated. In addition, the termination criteria and the parameters in the line search etc. should be specified.
Section II: Constrained convex optimization. 优化方法代写
(i)Adescription of the projected gradient method (Pg. 175, Eq (9.15)).
(ii)Thisdescription should contain a short description of the method and provide state- ments regarding both the convergence of the iterates and the associated rate of con- vergence (if any).
(iii)Finally,this algorithm should be applied to three problems: 4(i), 8.4(ii), and 8.5. The report should contain a log of iterations where f (xk) and xk ΠC (xk γ xf (xk)) (for any γ > 0) are tabulated. In addition, the termination criteria and the parameters in the line search etc. should be specified.优化方法代写
Submission deadline: December 13th.
IE 468：优化方法和模型（2019年春季）
专案（40％）Uday V.Shangbhag
项目要求：该项目是IE 468不可或缺的一部分，着重于在连续优化的背景下开发某些优化算法。本课程将包括提供一份包含以下内容的类型化报告（在附录中包含m文件和迭代日志）。
报告大纲。优化方法代写
第一节：无约束优化。
（i）四种算法中每一种的描述：（a）带有精确线搜索的梯度方法； （b）带有回溯线搜索的梯度方法； （c）具有精确/回溯线搜索的对角比例梯度法； （d）混合牛顿法。 （大多数已经完成）
（ii）该描述应包含对该方法的简短描述，并提供有关迭代的收敛和相关的收敛速率（如果有）的说明。
（iii）最后，每种算法都应应用于练习4.3（二次最小化）和示例5.8（Rosenbrock函数），并且报告应包含迭代记录，其中将f（xk）和xf（xk）制成表格。另外，应指定终止条件和线路搜索等中的参数。
第二节：约束凸优化。
（i）投影梯度法的说明（第175页，等式（9.15））。
（ii）该描述应包含对该方法的简短描述，并提供有关迭代的收敛和相关的收敛速率（如果有）的说明。优化方法代写
（iii）最后，该算法应应用于三个问题：8.4（i），8.4（ii）和8.5。该报告应包含一个迭代日志，其中将f（xk）和xkΠC（xkγxf（xk））（对于任何γ> 0）制成表格。另外，应指定终止条件和线路搜索等中的参数。
提交截止日期：12月13日。
其他代写：algorithm代写 analysis代写 app代写 加拿大代写 jupyter notebook代assembly代写 assignment代写 C++代写 code代写 course代写 dataset代写 北美作业代写 编程代写 考试助攻 source code代写 dataset代写 金融经济统计代写 java代写 web代写
