Introduction to Linear

Programming
-The Theory of the Simplex Method
Hongying FEI
[email protected]
Symbols used in a linear programming model
 The most common type of application of linear programming involves
allocating resources to activities. The amount available of each resource is
limited, so a careful allocation of resources to activities must be made.
= value of overall measure of performance

= level of activity for

= increase in that would result from each unit increase in level of activity

= amount of resource i that is available for allocation to activities for

= amount of resource consumed by each unit of activity

Data Needed
Data needed for a linear programming model involving the allocation of
resources to activities

Resources Usage per

Unit of Activity Amount of Resource
Resource Activities Available
1 2

1
2 to Z per
Contribution
unit of activity
m
Contribution to Z per
unit of activity
Standard Form of a linear Programming Model
 

s.t.
Recall: WYNDOR GLASS CO.
Production TimeActivities
per
Batch, Hours Production Time
Plant Product Available per week,
  Decision
Hours variables:
1 2

   

1 1 0 4
2 0 2 12
3 3 2 18
Profit per batch 𝑐  13=3 , 𝑐 2=5
5
 
s.t.
An optimal solution is feasible
solution that has the most favorable
value of the objective value.

The most favorable value is the

largest (smallest) value if the
objective function is to be maximized
(minimized).
  Slop-intercept form:
Optimal solution: (2,6)
Optimal value : Z= 36
 Terminology for solution of the Model
An corner-point feasible (CPF) solution is feasible solution that lies at a corner of
the feasible region.

For any linear programming problem with a

bounded feasible region, the problem must
possess CPF solutions and at least one optimal
solution. Furthermore, the best CPF solution
must be an optimal solution.

If a linear programming problem has exactly

one optimal solution, it MUST be a CPF
solution. If the problem has multiple optimal
solutions, AT LEAST two must be CPF
solutions!
Optimality Test
Consider any linear programming problem that possesses at least one optimal
solution. If a CPF solution has no adjacent CPF solutions that are better (as
measured by Z) ， then it MUST be an optimal solution.

Choose (0,0) as the initial CPF solution to examine, Z = 0

has better adjacent CPF solutions?

Choose the adjacent one (0,6) to examine, Z = 30

has better adjacent CPF solutions?

Choose the adjacent one (2,6) to examine, Z = 36

has better adjacent CPF solutions?

(2,6) is the optimal solution

 The Simplex Method
 Is for solving linear programming problems.
 Developed by George Dantzig in 1947 and applied by Leonid
Vitaliyevich Kantorovich to develop the theory of optimum allocation of
resources, who was awarded Nobel Prize 1975 in Economic Science
 Focuses solely on CPF solutions.
 Is an iterative algorithm.

(Checking whether any of adjacent CPF gives a positive rate of improvement in Z!)

(The current CPF solution is optimal!)

(Should Identify the rate of improvement in Z that would be obtained)

The Simplex Method
 The simplex method has proved to be a remarkably efficient method that
is used routinely to solve huge problems on today’s computers.

 The simplex method is always executed on a computer, and sophisticated

software packages are widely available.

It is necessary to translate the conceptually geometric procedure

in to a usable algebraic language of the simplex method.
Augmented Form of LP
 Objective function : Maximize

 Right-side : non-negative

 Decision variables : non-negative

 Functional constraints: equations

Convert Original Form to Augmented Form
 
Objective function : Maximize
 If objective function is 

Right-side : non-negative
 If certain right side is negative  multiple both sides by (-1).
 
Decision variables : non-negative
 For any , and replace by ,
 For unrestricted variable , replace it by the subtraction of two non-negative
variables, i.e. replace by ,
Convert Original Form to Augmented Form
 
Functional constraints: equations
 For any equivalent constraint , add slack variable to make
 For any equivalent constraint , subtract surplus variable to make
Example 1

  Original Form   Augmented Form

 Introduce
slack
Subject to variables
Subject to
and
 Example 2
  Original Form
  Convert Minimize Z to Maximize Z’ = -Z
 Introduce slack variables
 Introduce surplus variable
Subject to
 Set , replace by
 Set and replace by

  Augmented Form

Subject to

-
Exercises 2-1
Convert the following LPs to Augmented form.

 
Subject to
 
Subject to
Setting Up the Simplex Method

The simplex method is an The first step in setting up the simplex

algebraic procedure. method is to convert the functional
inequality constraints (except the non-
negativity constraints) to equivalent
The algebraic procedure is based equality constraints.
on solving systems of equations.

 
  Subject to
Subject to
Introducing slack variables

Original form Augmented form

Setting Up the Simplex Method
 Both forms of the model represent exactly the same problem.
 Augmented forms is much more convenient for algebraic manipulation
and for identification of CPF solutions.
 An augmented solution is a solution for the original decision variables
that has been augmented by the corresponding values of slack variables.

(3,2) (3,2,1,8,5)
 
  Subject to
Subject to
Introducing slack variables

Original form Augmented form

 Setting Up the Simplex Method
Degree of Freedom = number of variables – number of equations.

If the number of variables is larger than the number of equations, the degree of
Freedom indicates the number of variables that can be chosen to be set equal to
any arbitrary value (called non-basic variables) in order to solve the function
constraints (now equations) in terms of remaining the other variables (called
basic variables).
Basic Solution: All non-basic variables are set equal to zero and the basic variables to
be the solution of functional constraints (now equations).

If the basic variables satisfy the non-negativity constraints, the basic solution is a
basic feasible (BF) solution.

Basic Solution: An augmented corner-point solution.

Basic Feasible Solution: An augmented CPF solution.
 The Algebra of the Simplex Method
 Initialization :
Transform Original form to augmented form by introducing slack
variables when necessary.
   
Subject to Subject to

Introducing slack variables

 Set initial basic feasible (BF) solution :

Let all original variables be zero  the initial BF solution is (0,0,4,12,18).

  Optimality Test :
Objective function ( a l  
Increasing of both
t i m and can improve the
p
The current objective value Z= 3*0+5*0o= 0
ot objective function!
 The Algebra of the Simplex Method

 Determine
  the Direction of Movement :
 
5>3! Choose to increase!
Increase ? Rate of improvement in Z = 3.
( is called the entering basic
Increase ? Rate of improvement in Z = 5. variable )
  Solution Improvement (Determine how far to increase the entering basic
variable)
To increase as far as possible without leaving the feasible region while
keeping !

  Optimality Test :
Objective function (
The current objective value Z= 3*0+5*0 = 0
 The Algebra of the Simplex Method

 Determine
  the Direction of Movement :
 
5>3! Choose to increase!
Increase ? Rate of improvement in Z = 3.
( is called the entering basic
Increase ? Rate of improvement in Z = 5. variable )
  Solution Improvement (Determine how far to increase the entering basic
variable)
To increase as far as possible without leaving the feasible region while
keeping !

    
  No upper bound on

Minimum

Minimum ratio test

Minimum Ratio Test
The objective of the minimum ration test is to determine which
basic variable drops to zero first as the entering basic variable is
increased.

Decreasing the basic variable to zero will convert it to non-basic

variable for the next BF solution. Therefore, this variable is
called the leaving basic variable for the current iteration.
  is the leaving basic variable for the
current iteration

   (0,?,?,0,?)
Minimum Non-basic:
Basic :
Transformation
The objective of this step is to convert the equations to a more
convenient form for conducting optimality test with the new BF
solution.

   (0,?,?,0,?)
Minimum Non-basic:
Basic :
 Gauss-Jordan method of elimination
Transformation i.e. Gaussian Elimination

The objective of this step is to convert the equations to a more

convenient form for conducting optimality test with the new BF
solution.
(0) Z - =0  New BF solution
(0,?,?,0,?)
(1) =4
Non-basic:
(2) = 12 Basic :
(3) + =18
(3) =18

 Original BF solution

Non-basic:
Basic :
 Gauss-Jordan method of elimination
Transformation i.e. Gaussian Elimination

The objective of this step is to convert the equations to a more

convenient form for conducting optimality test with the new BF
solution.
(0) Z - =0  New BF solution
(0,?,?,0,?)
(1) =4
Non-basic:
(2) = 12 Basic :
(3) + =18
(3) =18

(0) Z + = 30
 Original BF solution
(1) =4
Non-basic:
(2) =6
Basic :
(3) - + =6
(3) =6
 Gauss-Jordan method of elimination
Transformation i.e. Gaussian Elimination

The objective of this step is to convert the equations to a more

convenient form for conducting optimality test with the new BF
solution.
(0) Z - =0  
New BF solution
(0,6,4,0,6)
(0,?,?,0,?)
(1) =4
Non-basic:
(2) = 12 Basic :
(3) + =18
(3) =18

(0) Z + = 30
 Original BF solution
(1) =4
Non-basic:
(2) =6
Basic :
(3) - + =6
(3) =6
 Optimality Test for the New BF solution
(0) Z + = 30 Z=30
 Original BF solution
(0,6,4,0,6)
(1) =4
Non-basic:
(2) =6
Basic :
(3) - + =6
(3) =6

  5   Increasing either or would result in moving

𝑧=30+3 𝑥 1 − 𝑥 4 towards one of the two adjacent BF solutions.
2

    The coefficient of is positive increasing would

lead to a better BF solution!
 
  Choosing to be the entering basic variable ! Z=36
     (2,6,2,0,0)
Non-basic:
Leaving basic variable Minimum Basic :
  𝑧=36 − 3 𝑥 − 𝑥
Transformation
4 5
2

 
Increasing either or would decrease
Z! Therefore, the current BF solution
(0) Z + = 30 must be optimal!

(1) =4
 New BF solution
(2) =6

(3) - + =6 Non-basic:
(3) =6
Basic :

(0) Z + + = 36

(1) =2
 Original BF solution
(2) =6
(0,6,4,0,6)
Non-basic: (3) + =2
Basic :
Question & Answer
Thanks !
Hongying [email protected]