Sunteți pe pagina 1din 54

m Introduction

-Theory
- Steps involved in formulation of LGPP
m Single Goal Programming
m Multiple Goal Programming
- Non premptive
- Premptive
¦ 

  

i Mathematical model similar to Linear Programming, however it


allows for multiple goals to be satisfied at the same time.

i Lpp(Linear programming problem) is based on the assumption that


the decision making has a single, qualifiable objective such as
maximization of Profit/output/sale or minimization of cost/ time.

i Often there are situations where instead of having single objective,


multiple objectives or goals may be set. In such situations we need
different technique that seek a compromise solution based on the
relative importance of each objective.

i Goal programming uses simplex method for finding the optimum


solution to a multi dimensional linear objective function with linear
constraints.
Goal programming has been widely applied to decision problems in
business organizations, government agencies, and nonprofit
institutions. Some applications includes ʹ

Accounting Analysis
Capital budgeting
Economic policy analysis
Energy resource planning
Facility layout and location
0 0  
  

Œ. V   

º
In this case only one goal has to be achieved

2.    



(a) Non-premptive goal programmingº In this case all the goals


are equally ranked & are of roughly comparable importance.

(b) Premptive goal programmingº In this case there is a hierarchy


of priority levels for the goals so that goals of primary importance
receive first priority attention, those secondary importance
receive second priority attention, and so forth(if there are more
than 2 priority levels)
 


V º Find out the decision variable of the major decisions


V º Formulate all the goals of the programme,these are
generally fixed by the desire of the decision maker or
resources or constraints
V º Express each goal in terms of constraints equations by
introducing negative and positive deviations
i.e DU or d- & d+ or Do
Where Du or d- = underachievement of goal
d+ or Do = over achievement of goal

V º Fix the priority levels of each goals (only in case of


multiple goals) & put these in achievement function
In general, once all goals & constraints are identified in a
problem, management should analyze each goal to see if
under or over-achievement of that goal is an acceptable
situation.

If management seeks to attain a goal met exactly both d- and


d+ must appear in an objective function.
Q A firm manufactures two products TV and DVD which must be
processed through two processes, Assembly & Finishing. Assembly
has 96 hours available and finishing has 78 hours available. For ΠTV
set requires 6 hours in Assembly & 4 hours in finishing while ΠDVD
set requires 5 hours in Assembly & 3 Hours in Finishing. If profit is Rs
90 per TV & Œ00 per DVD, find out the best combination of TV & DVD
to realize a maximum profit of Rs 2500.

Ô 
  

 6 4 90

 5 3 Œ00

 96 78 2500
Let XΠ& X2 be the no. of TV & DVD.
The given problem can be first formulated as LPP

Max. Z = 90 XŒ + Œ00 X2
Sub . to 6 XŒ + 5 X2 ч 96 - Assembly constraints
4 XŒ + 3 X2 ч 78 - Finishing constraints
Where XŒ & X2 ш 0

In case of goal programming, any goal stated can be º-


m Exactly met
m Under achieved
m Over achieved

Here the goal has to be set to as to make a profit of Rs 2500

Thus now the requirement can be expressed as


90 XŒ + Œ00 X2 = 2500
m Now using Deviational Variables ..
Let d- be the under achievement of profit goal & d+ be the over
achievement of profit goal.

The given problem can be expressed as the following Goal


Programming problem.

Minimize Z = d- (under achievement of profit goal)

Sub. toº 90 XŒ + Œ00 X2 + (d-) ʹ (d+) = 2500 (Goal constraint)

6 XŒ + 5 X2 ч 96 (Assembly constraint)
4 XŒ + 3 X2 ч 78 (Finishing constraint)
Where XŒ , X2, d+ & d- ш 0

We will now convert this into standard minimization problem.


Minimize Z = 0XΠ+ 0X2 + 0SΠ+ 0S2 +(d-) + 0(d+)
Sub. to 90 XŒ + Œ00 X2 + (d-) ʹ (d+) = 2500
6XΠ+ 5X2 + SΠ+ 0S2 = 96
4XΠ+ 3X2 + 0SΠ+ S2 = 78
Where XŒ, X2, SŒ, S2, (d-) & (d+) ш 0
Ë 
      

   

 Ë  "  " #$ #%  
Ë
! Ë |

#$ 90 Œ00 0 0 Œ -Œ 2500 Ñ

  6 5 Œ 0 0 0 96 Ñ

"  4 3 0 Œ 0 0 78 Ñ
o 90 Œ00 0 0 Œ -Œ "&

$ o -90 -Œ00 0 0 0 Œ
|
Ë       

   
Ë  "  " #$ #%

!

Ë

#$ -30 0 -20 0 Π-Π580

"  6/5 Œ Œ/5 0 0 0 96/5

"  2/5 0 -3/5 Œ 0 0 Œ02/5

o -30 0 -20 0 Π-Π&'

$ o 30 0 20 Œ 0 Œ

Since Cj ʹ Zj ш 0, thus solution is optimal


XΠ= 0, X2 = 96/5 , d- = 580 which means, target profit of 2500 is under
achieved by Rs 580 i.e the profit actually earned is Rs 2500 ʹ 580 = Œ920
Example 7.ŒŒ(N.D Vohra)
m A firm is engaged in producing two products , A
and B. Each unit of product A requires 4 kg of raw
material and 2 labor hours for processing,
whereas each unit of product B requires 3 kg of
raw material and 3 hours of labor , of the same
type. Every week ,the firm has an availability of
60 kg raw material and 96 labor hours .One unit
of product A sold yields Rs 40 and one unit of
product B sold yields Rs 35 as profit. find out the
best combination to realize a maximum profit of
Rs.Œ400.
O |V        O 

A 2 4 Kg Rs. 40

B 3 3 Kg Rs. 35

Constraints 60 Hours 96 Kg

The LPP equations would beº


Maximize Z = 40xŒ+ 35x2 (Profit)

Subject to 2xŒ + 3x2 ч 60 (Labor hours constraint)

4xŒ + 3x2 ч 96 (Raw material constraint)

xŒ, x2 ш 0
m The given problem can be expressed as the
following uO  

 

m Minimize Z =d-
Subject to 2xŒ + 3x2 ч 60
4xŒ + 3x2 ч 96
40xŒ + 35 x2 +(d-) ʹ (d+) = Œ400
Where XŒ , X2, d+ & d- ш 0

Now solving it via a Simplex Solution problemº-


Minimize Z = 0XŒ + 0X2 + 0SŒ + 0S2 +Œ(d-) + 0(d+)

Subject to 2xΠ+ 3x2 + SΠ+ 0S2 = 60


4xΠ+ 3x2 +0SΠ+ S2 = 96
40xŒ + 35 x2 +(d-) ʹ (d+) + SŒ + S2 = Œ400
Where XŒ, X2, SŒ, S2, (d-) & (d+) ш 0

Now solving it with help of Simplex Tables.


Simplex Tableau Œ
Ë       
   
 Ë  "  " #$ #%   
Ë
!
  Ñ     60 
"       96 Ñ
#$      Œ400 
o 40 35 0 0 Œ -Œ

$ o -40 -35 0 0 0 Œ
Simplex Tableau 2
Ë       
   

Ë    
Ë
! "  " #$ #%

   Ñ   Œ2 å

       24 Ñ
#$      440 åå
o 0 5 0 -Œ0 Œ -Œ

$ o 0 -5 0 Œ0 0 Œ
Simplex Tableau 3
Ë       
  

Ë  
Ë
! "  " #$ #%
"   Ñ     8

    Ñ Ñ   Œ8

#$     Ñ   400

o 0 0   -25/3 Œ -Œ

$ o 0 0 Œ0/3 25/3 0 Œ
m So , the solution is
m xŒ =Œ8
m X2 = 8
m d+ = 0
m d- = 400 , that means the profit goal is under
achieved by Rs. 400 and the maximum profit is
Rs. Œ000.
GOAL PROGRAMING WITH
MULTIPLE GOALS
Œ.Non-preemptive Goal programming
2.Preemptive Goal programming
@ 
 uO  


Assumes the decision-maker has a linear utility function with
respect to the objectives
Differential weights are used to various goals in line with their
relative significance
O 
 uO  


Uses prioritized goals
When optimal solution with respect to the higher priority goal is
obtained only then the next priority level is considered
NON-PREEMPTIVE GOAL
PROGRAMMING
ƒ In non-preemptive goal programming, we first
establish a goal for each objective and then seek a
solution that minimizes the sum of deviations of
these objectives from their respective goals

ƒThe goal programming assumes that the decision-


maker has I linear utility function with respect to the
objectives, that is to say, the marginal rate of
substitution between the objectives is linear,
regardless of the extent of deviation from the goal.
ƒThe deviations may be given different weights , called
m    
in accordance with the relative
significance of the objectives, and the solution
sought may be the one which minimizes the
weighted sum of the derivatives.
ƒthe weight used in a goal programming model are
indicative of the decision-maker͛s utility for the
various objectives.
ƒSpecifically, they measure the marginal rate of
substitution between objectives and the degree of
importance in attaining the goal of each objective
relative to the others.
EXAMPLE 7.Œ2 N.D VOHRA

The production manager of a company wants to schedule a weeks


production run for two products PΠand P2, each of which requires the
labour and materials as shown below.

PRODUCT
O O
Labour hours 2 4
Material MΠ4 5
(kg)
Material M2 5 4
(kg)

The weekly availability of resources is limited to 600 labour hours, Œ000 kg of


material MŒ and Œ200 kg of M2. The unit profit for PŒ and P2 is Rs.20 and Rs
32 respectively.

Products PΠand P2 are in fact new models and are replacements of the older
ones which have been discontinued very recently .
The manager would like to maximise the profit but he is equally
concerned with maintaining workforce of the division at the
nearly constant level in the interest of employee morale. The
workforce which consists of people engaged in production,
sales, distribution, peons and other general staff, consisted of
Œ08 persons in all. From a detailed study, it is known that
production of one unit of PΠwould maintain 0.3 person in the
workforce while one unit of P2 would maintain 0.75 person.
SOLUTIONº-

If we let xΠand x2 represent the number


of units of PΠand P2 respectively, respectively, to
be produced every week, the goals and
constraints of the problem cab be stand as
followsº

20xŒ+ 32x2 = 5400 goal-Œ


0.3xŒ+0.75x2 = Œ08 goal-2
2xΠ+ 4x2 <= 600 labour
4xŒ + 5x2 <= Œ000 material MŒ
5xŒ + 4x2 <= Œ200 material M2
This problem may not have a feasible solution which satisfies
both the goals. Further, to be able to solve this problem using
simplex method, we need to introduce deviational variables in
each of the constraints involving goals. They may be defined as
followsº

Let dŒ- = number of rupees below the profit goal of Rs 5400


dŒ+ = number of rupees above the profit goal of Rs 5400
d2- = number of people below the workforce goal of Œ08
d2+ = number of people above the workforce goal of Œ08

now, considering the nature of goals presumed that


overachievement of either of them would attract no penalty,
and we may seek a solution which would minimize the
underachievement of both of them. Accordingly, the problem
may be stated as follows
MINIMIZE Z = (dŒ-) + (d2-)

Subject to

20xŒ + 32x2 ʹ (dŒ+) + (dŒ-) = 5400


0.3xŒ + 0.75x2 + (d2-) - (d2+) = Œ08
2xΠ+ 4x2 + SΠ= 600
4xŒ + 5x2 + S2 = Œ000
5xŒ + 4x2 +S3 =Œ200
xŒ,x2,SŒ,S2,S3,dŒ-,dŒ+,d2-and d2+ ш 0
SIMPLEX TABLE-Œ

Ë V       V V V   
V    
             !"#

  $ %&        ! 
V            
V            

V            

         

         
' $ 
 $
SIMPLEX TABLE-2

Ë V       V V V   
V  
   $     !    #) ))$
( !$ $

  $    $ $     
V  $      $     )$
 $ &
V       $    ! 
$ 
V  #$      $     #
 $ 
         

     $    
'  $  !
$
SIMPLEX TABLE-3

Ë V       V V V   
V  
        !    
(
  *      +    
  $      $   )$ 
 & 
V  $          $
$

V            
         

       !  
'
SIMPLEX TABLE-4
Ë V       V V V Ë  
V  
       $      #$
(    ! 
$&
      $      !
$ 

       $ $    
$  
V            !

V            $
# 
$
         

        !  
' #$ $

SIMPLEX TABLE-5
Ë V       V V V Ë
V   
 
    $        #$
( $  #$ 
!
    $! $!   $   #

    + $      

V    $! $!   #$   


V    $ %      

         

   #$ $    #$  


'  !
The optimal solution to the problem as obtained from simplex
is xŒ =Œ50, x2 = 75, d2- = 27/4, s2 = 25, s3 =Œ50, with the
objective function value equal to 6.75 .

From this it is evident that workforce shall be Œ0Œ.25, with the


employment goal being underachieved to the extent of 6.75
people, while Œ50 kg of material M2 would remain unutilized.
The other variables are non-basic so that all the available
labour hours shall be used and the profit goal be met exactly.
DIFFERENTIAL WEIGHTING
In this case to avoid the problems associated with having large magnitudes
of difference, we may standardize the weights so as to seek minimizing the
percentage deviation from the goals in each objective.

For dŒ- º Œ/5400 and for d2- º Œ/Œ08

Further, if the decision-maker feels that achieving the profit goal is twice as
important as employment goal, he would give the weight as follows.

For dŒ-º 2x(Œ/5400)= Œ/2700 and d2- º Œ/Œ08

Here we will do managerial assessment based on the trade of between


different goals.
It is given that one person underachieved of the unemployment goal is
deemed equal to RsŒ00 in profit.
By this we may rewrite the objective function as

', 
SIMPLEX TABLE-Œ

Ë V       V V V   
V    
             !"#

  $ %&        ! 
V            
V            

V            

         

         


' #
SIMPLEX TABLE-2

Ë V       V V V   
V  
   $     !    #) ))$
( !$ $

  $    $ $     
V  $      $     )$
 $ &
V       $    ! 
$ 
V  #$      $     #
 $ 
         

     !$    
'  $  !
$
SIMPLEX TABLE-3

Ë V       V V V   
V  
        !    
(
  *      +    
  $      $   )$ 
 & 
V  $          $
$

V            
         

       !  
'
SIMPLEX TABLE-4
Ë V       V V V Ë
V  
       $     
(    ! 
$&
      $     
$ 

       $ $   
$  
V           

V           
# 
$
         

     $   !  
'  $
It is given that from the optimal solution we have xΠ=60 and
x2 =Œ20 , for a total profit 0f 20x 60 + 32x Œ20 = Rs 5040 , and
dŒ- = 360, implying an underachievement of the profit goal of
Rs 5400 by Rs 360. however, since d2- is equal to zero, it has the
importance that the employment goal of Œ08 would be met . In
terms of solution, the labour hours will be fully utilized while
the raw materials would remain unutilized º Œ60 Kg of material
MΠand 420 Kg of material M2.
QUESº-A firm producing two products radio and transistor which
must be produced through two process I and Process II.90 hours
are available in Process I & 72 hours are available in Process II.
One radio requires 6 hours in Process I and 3 hours in process II.
One transistor requires 3 hours in Process I and 6 hours in Process
II. The profit is Rs.Œ20 per radio and Rs 90 per transistor. Find the
optimum solution in GPP, the firm has set up the following goals
and has assigned them priorities PŒ,P2 & P3 where PŒ is most
important. Find the optimum solution.

O  | u 
PŒ Production of radio goal of Œ3.
P2 To achieve profit goal of
Rs.Œ950
P3 Production of transistor goal
of 5.
Let Dur = Under- achievement of radio goal
Dor = Over- achievement of Radio goal
Dup = Under-achievement of Profit goal
Dop = Over-achievement of Profit goal
Dut = Under-achievement of transistor goal
Dot = Over-achievement of transistor goal

Objective function will be,

Minimize Z= PΠDur +P2Dup + P3Dut


Subject Toº- Œ20 XŒ + 90X2 + Dup-Dop ( Profit Goal)
XŒ + Dur ʹDor =Œ3 (Radio Goal)
X2 + Dut ʹ Dot =5 (Transistor Goal)
6XŒ+3X2+SŒ+0S2=90 (Process I constraint)
3XŒ + 6X2+0SŒ+S2=72 (Process II constraint)

Where XŒ,X2,SŒ,S2,Dur,Dor,Dup,Dop,Dut,Dot ш0
PΠis more important than P2, it means P2 goal will not be
achieved until PΠgoal has been achieved. Similarly P3 goal will
not be achieved till P2 goal is achieved.

Putting the above problem in the Simplex Tableº-


SIMPLEX TABLE-Œ
OOO
Ë V   V V        


PŒ Dur Œ3 Œ 0 0 0 0 0 Œ -Œ 0 0 Œ3
P2 DuP Œ95 Œ20 90 0 0 Œ -Œ 0 0 0 0 65/
0 4
P3 Dut 5 0 Π0 0 0 0 0 0 Π-Π
0 SŒ 90 6 3 Œ 0 0 0 0 0 0 0 Œ5
0 S2 72 3 6 0 Π0 0 0 0 0 0 24
Cj- P3 0 -Œ - - - 0 - 0 - Œ
Zj
P2 - - - - - Π- 0 - 0
Œ20 90
PΠ-Π0 - - - 0 - Π- 0
CŒ-ZŒ º-

PŒ= 0-Œ = -Œ
P2= 0-Œ20= -Œ20
P3= 0-0 = 0

Similarly we have calculated the values for PŒ, P2,P3.

' Each of the priorities PŒ, P2 and P3 has separate Zj and Cj-Zj in the
order of their priority of goals. It is because the goals do not have the
same dimension and cannot be just added.

'Starting with the most important priority say PŒ, the most negative
Cj-Zj is selected and entered. If there is no negative value in that row,
we move to the next most important priority P2 and work the most
negative Cj-Zj value.
SIMPLEX TABLE 2
OOO
Ë V   V V        


             
O  O )  )         
O             -
 V         &   
 V           
 O          
'
O          
)
O          
SIMPLEX TABLE 3
OOO
Ë V   V V        

    * $        

O  O            

O             
    $        
*
 V #  )$        
 $

' O          

O          

O          
SIMPLEX TABLE 4
OOO
Ë V   V V       

            
O  O            

O     $       

     $       

 V )       ) )  

' O   $       

O           
O          
There still exist -2 and -60 below Dur but it is to b disregarded as it will
again lead to ass under achievement of radio goal.

So XŒ=Œ3
X2=4

From the above table it is concluded thatº-

' The first goal PŒ to produce Œ3 radio (goal) XŒ=Œ3 has been achieved.

'The second important goal P2= Rs Œ950 has been under achieved by Rs
30 i.e. only
Rs (Œ950-30) = Rs Œ920 has been achieved.
(Œ20 3 + 90 4) =Rs Œ920

' The third goal P3 to produce 5 transistors is underachieved by Œ. We


still have 9 hours available in process II to manufacture Πtransistor but
we have no time left in process I, so it cannot be produced.
Example 7.Œ3 (N D Vohra)
m A firm is engaged in producing two products A and B
, which yield a unit profit of Rs. 400 and Rs.
300,resp.Both of these require a certain raw
material whose supply is limited to 300 kg per
month. The raw material requirement is one kg per
unit of each product. Further , it is known that one
unit of A requires 2 hours of processing while each
unit of B needs one hour. The firm has a normal
processing capacity of 400 hrs per month. The
marketing department has stated that according to
the prevailing conditions , and likely to continue in
the forseeable future , the maximum number of
units of A and B which can be sold each month is
Œ50 and 350 , resp.
The manager of the firm has set the following goals
in order of their importance ʹ the priorities being
preemptive.
m GOAL μ Avoid under-utilisation of normal
productive capacity.
m GOAL 2º Sell the maximum no. of units of products
A& B .However ,since the two products yield the
ratio of 400º300,the manager would give the
weightage to the sales of the products in the same
ratio.
m GOAL 3ºMinimise overtime of processing capacity.
You are required to formulate this as goal
programming problem and solve it graphically.
m Let xΠand x2 be the number of units of products A & B resp. to be produced
every month. O |V    O  O / 
m Now formulating      "V
A(xŒ) 2 Œ Kg Rs. Œ50 units
this as goal 400
programming B(x2) ΠΠKg Rs. 350 units
300
problemº-
Constraints 400 300 Kg
m   . Hours
(Raw material constraint)
    , (Processing capacity constraint)
Where dŒ- = no. of hours unutilized processing capacity
dŒ+ = no. of hours overtime processing operation
Since the sales goals given are in terms of the max. possible sales
volumes ,they an be expressed as º-
 ,  ,
(Here we are taking only deviational variables for under achievement and not for
overachievement as the sales goals are the maximum ones.)
m With PΠ, P2 , P3 being the preemptive priority
factors , the goal programming problem may be
stated as followsº
m Minimize Z = PŒdŒ +P2(4d2+3d3)+ P3dŒ
m Subject to
XŒ+ X2 ч 300
2XΠ+ X2 + dΠdΠ= 400
XŒ + d2 = Œ50
X2 + d3 = 350
XŒ , X2 , dŒ 0 d2 0d3 ш 0
Thank You͙

S-ar putea să vă placă și