Fall 2017 -- STOR 612
Deterministic models in OR
Instructor: Gabor Pataki.
TA: Weibin Mo
Class : Tue, Thu 5.00-6.15 PM, Hanes 130
Office hours :
For Gabor: Wednesday 3.30-5.00 Pm, Hanes 307.
Usually I am also available after class.
For Weibin: Wednesday 2-3:30 pm, Hanes B40
First midterm: Oct 3, Tue, during classtime, Hanes 130. In class, closed book, closed notes. Subject: Only LP formulations.
Exam 1 in 2014.
Exam 1 in 2014 with partial solutions.
Exam 1 in 2015
Exam 1 in 2015 with partial solutions.
Exam 1 in 2016
Exam 1 in 2016 with solutions.
Exam 1 in 2017
Exam 1 in 2017 with solutions.
Second midterm: Nov 14, Tue, in class time. I will give some extra time at the end.
In class, closed book, closed notes. Subject: shortest paths, and dynamic programming.
We call Category 1 students: PhD students in OR, Stat, INSTORE, KFBS, Math, CS.
Everyone else is Category 2.
The exam will be different for Cat1 and Cat2 students.
Cat1 students need to remember and reproduce algorithms exactly as covered in class, and reproduce their correctness proofs.
Cat 2 students will need to only state the algorithms, and theorems that were covered in class, but not the proof of their correctness. However: Cat 2 students will still need to **invent** some algorithms, and prove correctness.
Eg. Cat 1 students may be asked to state Dijkstra's algorithm, and prove its correctness. Cat 2 students may be asked only to state Dijkstra's algorithm.
Still, it is strongly advised for cat 2 students to study, and absorb all algorithms, and their proofs.
Some questions on the exam may be about some small modifications of these algorithms.
Exam 2 in 2014.
Exam 2 in 2014 with solutions.
Exam 2 in 2015.
Exam 2 in 2015 with solutions.
Exam 2 in 2016.
Exam 2 in 2016 with solutions.
Final: Tuesday, dec 12, 4 PM. In class. The final exam is cumulative.
It is on the same model as exam 2, (different for MS/undergrad, and PhD, etc.)
Final in 2014
Final in 2014 with partial solutions
Final in 2015
Final in 2015 with solutions
Final in 2016
Final in 2016 with partial solutions
aug 20. Syllabus and homework policies (pdf file)
- aug 20. HW is due BEFORE the beginning of class, or it can be placed in the instructor's
mailbox before the beginning of the class. Late hws cannot be accepted.
- aug 20.
How to submit homeworks with an AMPL ingredient: you need to
- Submit a handwritten solution explaining the LP model as usual, and in addition
a printout of your files, and of the run, and
send an email containing the model and data files to the TA, Weibin Mo
whose address is "harrymok" (don't ask) "at" "email" "dot" "unc.edu".
As part of the grading, he will run all of them to check their correctness.
- aug 20.
Generalities about AMPL hws:
I usually give the optimal solution value, so you can check your work.
If you get an objective value different from the one given, then it is very likely
that your model is incorrect. The data was usually made up in a way, so that incorrect models
will give a solution value different from the correct model's.
- Pretest results. The pretest The pretest with solutions
Q1 and Q2 were worth 2 pts, the rest of the questions 1 pt each.
IF: your grade is 4 or below, probaby you should drop the course, as your preparation does not seem to be sufficient. If you still decide to stay, you will need to spend a lot of time on catching up on proofs and discrete math. IF your grade between 5 and 6, then you will need to spend a lot of time catching up. If your grade is >=7, then you are ready to take the course. It is still a good idea to spend time on proofs and discrete math.
How to catch up: I recommend: 1) Discrete Mathematics and Its Applications, by Rosen (this should be easy) 2) Discrete Mathematics: Elementary and Beyond by Lovasz, Pelikan and Vesztergombi (this should be harder).
Linear Programming handouts
Networks and DP handouts