Semester.ly

Johns Hopkins University | EN.550.661

Foundations of Optimization

3.0

credits

Average Course Rating

(4.53)

This course considers algorithms for solving various important nonlinear optimization problems and, in parallel, develops the supporting theory. Primary focus will be on unconstrained and bound-constrained optimization. Topics will include: necessary and sufficient optimality conditions; gradient, Newton, and quasi-Newton based line-search and trust-region methods; linear and nonlinear least-squares problems; linear and nonlinear conjugate gradient methods; stochastic optimization; optimal gradient methods; structured non-smooth optimization, and derivative-free optimization. Special attention will be paid to the large-scale case and will include topics such as limited-memory quasi-Newton methods, projected gradient methods, and subspace accelerated two-phase methods for bound-constrained optimization. Recommended Course Background: Multivariable Calculus, Linear Algebra, Real Analysis such as AS.110.405

Fall 2012

(4.56)

Fall 2013

(4.62)

Fall 2014

(4.41)

Fall 2012

Professor: Daniel Robinson

(4.56)

Students said the professor did a good job of explaining concepts and the professor, alongside the TA were very accessible for extra help. They also liked the mix of theory and coding on the homework assignments. Suggestions for improvement included more applications and projects to practice skills. Some students wished the class had been less theoretical and more practical. Prospective students should have a background in linear algebra and matrix analysis.

Fall 2013

Professor: Daniel Robinson

(4.62)

Students widely touted the instructor as one of the best aspects of the course, saying he could, “explain complex things in an extremely clear and beautiful way.” Issues with the course varied greatly among students, but several found the course chal enging. Suggestions for improvement included providing more examples or tougher homework so that students could determine if they had mastered the course material. Prospective students should have a good familiarity with MATLAB, and that they should know how to develop a proof before taking this course.

Fall 2014

Professor: Daniel Robinson

(4.41)

Students praised this course and its instructor for focusing equally on theory and practical application of concepts. Perceived issues with the course varied. Multiple students disliked that the course seemed to move quickly, leaving little time for those without basics skills to catch up. In addition, multiple students disliked that information was taught using PowerPoint slides. Suggestions for improvement included a desire by multiple students that the course provide students with more example and homework problems in order to reinforce concepts. Prospective students should know that students broadly believed that having a strong background in linear algebra was helpful when taking this course.