You are on page 1of 33

IEEN 5335

Principles of Optimization

Lecture 1

Dr. Joon-Yeoul Oh

Dept. of Mechanical & Industrial Engineering


Texas A&M University – Kingsville
Some Internet resources
 Textbook web site: this is for 9th edition, which provides
more free stuff.
http://highered.mheducation.com/sites/0073376299/student
_view0/informs_articles.html
 Optimization, Frequently Asked Questions (FAQ)
http://www.mcs.anl.gov/otc/guide/faq/linear-
programming-faq.html
 Mathematical Programming Glossary
http://glossary.computing.society.informs.org/
 Institute for Operations Research and Management Science
http://www.informs.org/

2
LINDO

 http://www.lindo.com/index.php?option=com_content&vie
w=article&id=34&Itemid=15

 For downloading, click Download Classic LINDO

 Please download LINDO in your computer. No


worry…it’s free!

3
Why learn OR?

 Humans are natural optimizers


 You can be a hero: save your company time, money,
resources.
 Many companies use it to solve operating problems.
 Software is widely available (Excel, etc)
 It’s easy to do “what if” analyses.

4
OR background

 Operations Research (OR) started just before World War II


in Britain with the establishment of teams of scientists to
study the strategic and tactical problems involved in
military operations. – computer limitation
 The objective was to find the most effective utilization of
limited military resources by the use of quantitative
techniques.
 Following the war, numerous peacetime applications (post
war economy) emerged, leading to the use of OR and
management science in many industries and occupations. –
better computers

5
What is Operations Research?

 Operations Research (OR) is the study of mathematical


models for complex organizational systems.
 OR uses mathematical relationships and numerical data
 Optimization is a branch of OR which uses mathematical
techniques such as linear and nonlinear programming to
derive values for system variables that will optimize
performance.

6
7
See textbook table 1.1
OR terms
 Model
 Mathematical representation of a real world
 Abstraction of reality
 Small-scale representation of a large object
 System: A functionally related group of elements, especially:
 The human body: the nervous system; the skeletal system, etc
 A group of interacting mechanical or electrical components.
 A network of structures and channels, as for communication,
travel, or distribution.
 A network of related computer software, hardware, and data
transmission devices.
 Etc.
8
Phases of an OR study

 The principal phases for implementing OR in practice


include:

1. Defining the problem and gathering data


2. Formulating a mathematical model
3. Deriving solution(s) from the model
4. Validation of the model
5. Implementation of the solution

 This process is iterative, with feed-back between the steps

9
Phases of OR – validating the model

 How do you know it (the OR model) reflect reality?

 Validation efforts
 Recreating the past
 Expert opinion
 Start slowly and consider the end-user early

10
Where to put your effort?

 Different answers for students and practitioners; different


answers for different types of practitioners
 40% / 20% / 40% for defining the problem and
finding the data/crunching numbers/analysis,
packaging and presentation of results
 In the model building stage:
 15% to build the model; 85% to find, manage and
“clean-up” the data

11
OR basic concepts

 How to best use resources.


 Constraints,
 Cost or Value
 Model brings it together.
 Objective Function
 Constraint Equations

12
OR constraints

 Limitations on resources, e.g.


 We have 12 telephone lines.
 We must finish in two weeks.
 We can get 200 tons of steel.
 800 volunteer-hours of labor.

13
OR costs

 Unit Cost of resource, e.g.


 Telephone lines costs $10/hr.
 Saving a day costs $2000.
 Each axle takes 85 lb steel.
 Winterizing: 75 hrs labor.

14
OR values

 Benefit of a solution, e.g.


 Revenue avg: $30 / call.
 Being late costs $5000/day.
 Each axle is worth $145.
 Winterizing saves $400/yr.

15
OR model
 A mathematical model consists of:
 Decision Variables, Constraints, Objective Function,
Parameters and Data
 Linear program (LP)
 Integer program (IP)
 Network programming
 Nonlinear programming
 Etc.

16
OR Taxonomy

17
OR Taxonomy (cont.)

18
OR Taxonomy (cont.)

19
Mathematical Programming - terms

 A solution is an assignment of values to variables.


 A feasible solution is an assignment of values to variables
such that all the constraints are satisfied.
 The objective function value of a solution is obtained by
evaluating the objective function at the given solution.
 An optimal solution (for minimization) is one whose
objective function value is less than or equal to that of all
other feasible solutions.
 For maximization, the optimal solution is one whose the
object value is greater than or equal to that of all other
feasible solutions.
20
OR – simple Transportation example
Distribution centers Stores

21
OR – simple Transportation model

22
OR – Queueing model

23
OR – general LP form

24
Characteristics of LPs

 Decision variables (DVs) are not negative (non-


negativity)
 Criterion for choosing the best values for the DVs is a
linear function (first powers and no cross-products)
 Resource restrictions are a set of linear equations or
inequalities.

25
Formulating Linear Programs

 Translating a problem into a linear program is part art and


part science.
 “Typical” examples:
 Product mix
 Advertising media selection
 Workforce scheduling
 Job Training
 Gasoline blending

We will solve these examples later.

26
LP - example
 Q: How many chairs and/or tables to make?
 Each require pine, oak & time
 Limits on pine, oak and time.
 Profit: $6/chair, $12/table.
 Data
Item Chair Table Limit
Profit $6 $12 ---
Pine 3 bf 12 bf 120 bf
Oak 8 bf 4 bf 160 bf
Time 6 hr 10 hr 150 hr
27
Chair and Table example

 Decision Variables (x)


 Use to make the decision
 Objective function is z(x)
 Constraints are gi(x)
 Here, x = (C,T)
 C = # chairs, T = # tables.

28
Chair and Table example - model

 Objective Function  LINDO format


Max: z = 6 C + 12 T
 Constraints Max 6 C + 12 T
s.t.
Pine: g1 = 3 C + 12 T ≤ 120
3 C + 12 T <= 120
Oak: g2 = 8 C + 4 T ≤ 160
8 C + 4 T <= 160
Time: g3 = 6 C + 10 T ≤ 150 6 C + 10 T <= 150
Non-neg: C ≥ 0, T ≥ 0 end

29
Chair and Table example – trial solution

 Let C = 10, T = 5
 z = 6 × 10 + 12 × 5 = 120
 Constraints
 Pine: 3 × 10 + 12 × 5 = 90 < 120
 Oak: 8 × 10 + 4 × 5 = 100 < 160
 Time: 6 × 10 + 10 × 5 = 110 < 150
 And, Non-neg: C > 0, T > 0

30
C & T example – trial solution (cont.)

 What we know:
 (C, T) = (10, 5) is feasible
 z(10, 5) = $120
 What we do not know
 Whether z(C, T) > $120 for some feasible (C, T)
 We also know …
 30 bf of pine left
 60 bf of oak left
 40 hr of time left
 We could make more!
31
C & T example – trial 2

 Make Only Tables !!!


 Pine: 120/12 = 10 T ≤ 10
 Oak: 160/4 = 40 T ≤ 10
 Time: 150/10 = 15 T ≤ 10
 Profit: 10 × $12 Z = $120

32
C & T example – trial 3

 Make Only Chairs !!!


 Pine: 120/3 = 40 C ≤ 40
 Oak: 160/8 = 20 C ≤ 20
 Time: 150/6 = 25 C ≤ 20
 Profit: 20 × $6 Z = $120

33

You might also like