You are on page 1of 10

IEOR 4004: Introduction to

Operations Research
Deterministic Models
January 22, 2014
Syllabus
20% homework assignments 1st homework is
already available
40% midterm
on Courseworks
40% final exam
Lectures Monday, Wednesday 7:10pm-8:25pm
Recitations: Friday 12:30pm-2pm
Instructor: Juraj Stacho (myself)
office hours: Tuesday 1pm-2pm
Teaching assistant (TA): Itai Feigenbaum
office hours: Friday, after recitations 2:15pm-3:15pm
all parameters Summary goal is to
known minimize or
maximize
Goal of the course:
Learn foundations of mathematical modeling of
(deterministic) optimization problems
Linear programming problem formulation (2 weeks)
Solving LPs Simplex method (4 weeks)
Network problems (2 weeks)
Integer Programming (1.5 weeks)
Dynamic Programming (1.5 weeks)
Non-linear Programming (1 week time permitting)
2 weeks reserved for review (1 before midterm)
What this course is/is not about
Is about: Not about:
mathematical modeling coding (computer
(problem abstraction, programming)
simplification, model engineering (heuristics,
selection, solution) trade-offs, best practice)
algorithms stochastic problems
foundations of optimization (uncertainty, chance)
deterministic models solving problems on
example real-world models real-world data
typical model: medium- modeling risk, financial
term production/financial models, stock markets,
planning/scheduling strategic planning
Mathematical modeling
Simplified (idealized) formulation
Limitations
Only as good as our
assumptions/input data
Cannot make predictions beyond
the assumptions

We need
more maps
Mathematical modeling
Problem
simplification
Model
Model
formulation
Algorithm
selection
Problem Numerical
calculation

Solution
Interpretation
Sensitivity analysis
Model selection
trade-off between accuracy (predictive power)
and model simplicity (being able to solve it)
Predictive power
Low High
(quality of prediction)

Simple Model Complex


Easy Hard
to compute
Solving (if not impossible)
to compute
a solution Linear programming
a solution
(seconds) Network algorithms
Integer programming (lifetime of
Dynamic programming the universe)
Modeling optimization problems
Optimization problem Mathematical model
decisions decision variables
goal (objective) objective function
constraints constraint equations
Linear
Formulating program problem
an optimization
Linear
Decision variables objective x 1 , x2 , x3 , x 4
x1
Minimize 2x2 + 3x3 +
Objective (1 2x2)2

(x1)2 + (x2)2 + (x3)2 + (x4)2 2


Constraints
x1 2x2 + x3 3x4 1
x1 + 3x2 + 2x3 + x4 2
Domains 2x1x3 = 1

Linear x1 0 x3 in {0,1}
constraints x2 0.5 x4 in [0,1]

Sign restriction
Mathematical modeling
Deterministic = values known with certainty
Stochastic = involves chance, uncertainty
Linear, non-linear, convex, semi-definite

You might also like