Jonathan Eckstein's Business Decision Analytics under Uncertainty Classes
33:136:400:01/02
(Fall 2017)
This site is for both Fall 2017 sections of the course. The
Spring 2017 version and earlier versions are also available.
Announcements (As of
November 16, 2017 04:58 PM)

I am planning to postpone the second midterm to
November 20, because we are a class behind where I was hoping to
be in the syllabus. An amended
class schedule is available.

Practice material for the second midterm is now available on
Blackboard (with updates to correct typos 5:30pm November 15).

Office hours: There will be special preexam
office 1:303:30pm on Friday, November 17. There will be no office
hours on Thanksgiving week.

This website will be used for publicaccess materials. Materials
needing password protection will be on
Blackboard.
Usual Office Hour Schedule
My office is 100 Rockafeller Road, room 5145. Office hours are
tentatively scheduled for
 Tuesdays 1:304:30 pm from September 12 through December 12, except
 No office hours October 12 (after first midterm)
 No office hours November 21 (Thanksgiving week and after second
midterm)
 Extra office hours are typically scheduled before exams; check the
space above and class announcements.
This schedule is subject to change. Changes will be announced above. You may also make appointments if regularly scheduled
office hours conflict with your class schedule.
Handouts and Class Materials
 Wednesday, September 6: Introduction, procedures, syllabus,
introduction to decision making under uncertainty
 Monday, September 11: Introduction to decision trees, probability
review
 Wednesday, September 13: Conditional probability and Bayes' formula
 "Bayes' rule" and TB testing problem, pages 1314 of the textbook
 Material processing problem
(similar but not identical to page 69 of the textbook)
 Homework 1, due September
20. Please check your
email for an announcement regarding this assignment; in particular,
please substitute "is considering an option" for "has
acquired and option" in the first sentence of problem 1.
 Monday, September 18 (covered by substitute instructor): More
Bayes analysis and decision trees
 Wednesday, September 20 (covered by substitute instructor): More
Bayes analysis and decision trees, risk considerations
 The planned material on utility functions was skipped since the
substitute instructor covered material more slowly than anticipated
(this material is not covered in the exams)
 Homework 2, due September
27
 Monday, September 25: Critical fractile analysis, including "chunky"
demand
 Wednesday, September 27: Critical fractile case study
 Monday, October 2: Introduction to
dynamic programming
 Discuss format of upcoming exam
 Shortest path problem (book Example 3, page 220)
 Discussion of the computational efficiency of dynamic programming
 Wednesday, October 4: Review for exam

Monday, October 9: First midterm
exam (practice material on Blackboard)
 Wednesday, October
11: More dynamic
programming by hand
 Knapsack problem (on board)
 Homework 4, due October 18
(note: the data table heading should read "profit", not
"profit per flight")
 Monday, October 16: Review exam results,
dynamic programming by spreadsheet, start Python
 Inventory problem (Example 4, page 227 of
textbook)
 Wednesday, October 18: Python refresher/primer
 Monday, October 23: Our first dynamic
programming solution in Python (substitute instructor)
 Wednesday, October 25: More deterministic dynamic programming in
Python
 Integer knapsack problems
 General algorithmic/code template for deterministic dynamic
programming
 Airplane flight allocation problem
 Homework 6, due November 1
 Monday, October 30: Introducing stochastic dynamic programming
 Probabilistic inventory example on pages 276279 of the book
 Python code solving this example
 Larger example
 Wednesday, November 1: Stochastic programming with shortage
penalties or net present value, start elementary stochastic processes
 Amended class schedule
 Modified inventory problem that has shortage costs instead of a requirement to
always be able to meet demand
 Including net present value and discounting in dynamic programming
 Bernoulli processes: Bernoulli, binomial, and geometric random
variables
 Homework 7, due November
8
 Monday, November 6: More elementary stochastic processes, using
Poisson and binomial distributions within dynamic programming calculations
 Wednesday, November 8: The curse of dimensionality,
introduction/refresher on Monte Carlo simulation
 Monday, November 13: Discuss midterm
format, more Monte Carlo simulation
 Wednesday, November 15: Review for exam, start
Monte Carlo simulation of queuing systems
 Exam review material is posted on
Blackboard (with typo corrections 5:30pm November 15)

Taxi
repair shop problem

Lecture notes on Monte Carlo queuing models
are posted on
Blackboard ("First notes on queuing: repair shop and horses")
 Monday,
November 20: Second midterm exam
Homework Solutions
All solutions will be posted on
Blackboard and require a
password. Look under "Course documents" and then the
"Homework solutions" folder. The files are in PDF format unless otherwise specified.
 The solution to homework 1 is posted on
Blackboard
 The solution to homework 2 is posted on
Blackboard
 The solution to homework 3 is posted on
Blackboard
 The solution to homework 4 is posted on
Blackboard
 The solution to homework 5 is posted on
Blackboard
 The solution to homework 6 is posted on
Blackboard
 The solution to homework 7 is posted on
Blackboard
 The solution to homework 8 is posted on
Blackboard