Ming Gu, Professor of Mathematics

Math 170, Mathematical Methods for Optimization, Fall 2010.

Class Information
- Sample Final Exam . This was the final exam in 2008.
- Sample Midterm Exam .
- The required textbook is
  - Joel Franklin Methods of Mathematical Economics, SIAM.
- Matlab is going to be the programming language for this class. Good matlab programming skills can go a long way to earn you a descent grade in this class. The Math Department offers a 1-credit class Math 98 this semester to help those who want to learn or enhance their matlab skills. This class lasts for about 5-6 weeks. If you would like to enroll or to know more about this class, please contact Boris Ettinger at ettinger@math.berkeley.edu or visit /~ettinger/math98.html .
- If you do not yet know it, you can ask your GSI or myself about the username and password on your computer account in the basement computer lab.
- Information about Instructor Prof. Ming Gu
  Office: Evans 861
  Office Hours: WF 2:00-3:00PM, or by appointment.
  Phone: 642-3145
  Email: mgu@math
More Details
- More general information about the course can be found at here .
Homework Assignments
- Term Project , Due Dec. 1, 2010.
- Homework Assignment #1 , Due Sept. 8, 2010 in class.
- Homework Assignment #2 , Due Sept. 15, 2010 in class.
- Homework Assignment #3 , Due Sept. 22, 2010 in class.
- Homework Assignment #4 , Due Sept. 29, 2010 in class.
- Homework Assignment #5 , Due Oct. 6, 2010 in class.
- Homework Assignment #6 , Due Oct. 13, 2010 in class.
- Homework Assignment #7 , Due Oct. 20, 2010 in class. Modified Oct. 16.
- Homework Assignment #8 , Due Oct. 27, 2010 in class.
- Homework Assignment #9 , Due Nov. 3, 2010 in class.
- Homework Assignment #10 , Due Nov. 10, 2010 in class.
- Homework Assignment #11 , Due Nov. 17, 2010 in class.
- Homework Assignment #12 , Due Nov. 24, 2010 in class.
- Homework Assignment #13 , Due Dec. 1, 2010 in class.
Matlab Codes
- This is a Matlab session to solve an LP. Download it, and try to run something similar to get familiar with the Simplex Method.
- This is an interactive Matlab program to solve an LP with a shortened Simplex Tableau. Download it along with functions SimplexTableau.m and getr.m to run the simplex method.
- This is another interactive Matlab program to solve an LP with Extended Simplex Tableau. Download it along with functions SimplexTableau.m and getr.m to run the simplex method based on updating the Simplex Tableau.
- This is the third interactive Matlab program to solve an LP with the Revised Simplex Tableau. Download it along with functions SimplexTableau.m and getr.m to run the simplex method. You should compare three versions in both the results (they should be almost identical) and the computations involving the Tableaus.
- This is an interactive Matlab program to solve two-person zero-sum game using the Revised Simplex Tableau. Download it along with functions RevisedSimplex.m and getr.m to give it a try.
- This is an ILP solver using the revised simplex tableau. Download it along with function ILPRevisedSimplex.m . Give it a try.
- This is a NetworkFlow solver based on the idea of unsaturated path. The solver is based on the algorithm in the book and hence is not very efficient, but it does get the job done for small to middle sized networks. Try it with a driver routine .
- This is a Optimal Assignment code based on Network Flows.
- This is a code for solving the Transportation Problem based on Network Flow code above.
- This is a code for solving the Quadratic Programming Problem based on Wolfe's Algorithm. You need the subroutines QPRevisedSimplex.m and Revisedgetr.m for the code to work.
- This is a sample LP solver using the revised simplex tableau for the canonical LP problem. Download it along with functions ILPRevisedSimplex.m , RevisedSimplexTableau.m , and Revisedgetr.m . You can base your project on these codes if you so choose.
- These are two unfinished codes BLPRevisedSimplex.m for doing the simplex method on the bounded LP and BURevisedgetr.m for changing the index when a variable from the upper bound has been pushed downward to reduce the objective. Again the codes are incomplete. You can choose to write your own code based on them.
Web Links