Vous êtes sur la page 1sur 10

Lecture  3:  

Program  Evaluation  and  


Review  Technique  (PERT)

©  J.  Christopher  Beck  2008 1


Outline
n Program  Evaluation  and  Review  
Technique  (PERT)

©  J.  Christopher  Beck 2


Readings

n P  Ch 4.3
n Some  slides
taken  from
Twente  University
(see  Pinedo  CD)

©  J.  Christopher  Beck  2015 3


Uncertain  Processing  Times
n Great,  scheduling  is  easy!
n In  the  real  world,  do  we  really  know  the  
duration  of  a  job?
n What  if  we  have  estimates  of  duration?
n What  if  we  have  a  distribution:
n pj =  (μj,  δj)?

©  J.  Christopher  Beck  2015 4


Program  Evaluation  &  Review  
Technique  (PERT)
n Idea:  estimate  pj and then  just  use  CPM  
to  estimate:
n Ê(Cmax)  – expected  makespan
n Ṽ(Cmax)  – variance  of  makespan
n It  is  pretty  simple!
n You  are  responsible  for  P.  4.3!
n It  is  also  pretty  ad-­hoc
n See  [Beck  &  Wilson  2007]
5
Estimating  (μj  ,  δj)
n Assume  you  have  3  estimates  of  pj
n Optimistic:  paj
n Most  likely:  pmj
n Pessimistic:  pbj “No  battle  plan
survives  the  first
n Reasonable  estimates: encounter  with  the
n μj =  (paj+4pmj+pbj)  /  6 enemy.”
n δj =  (pbj-­paj)  /  6

©  J.  Christopher  Beck 6


PERT  Steps
n 1.  Find  μj  ,  δj2
n i.e.,  using  estimates  on  previous  slide
n 2.  Use  CPM  to  find  critical  path(s)
n with  pj =  μj
n 3.  Estimated  expected  value  and  
variance  of  Cmax Eˆ (Cmax ) = ∑ µˆ j
n Assume  makespan is j∈J cp

normally  distributed Vˆ (C ) = 2
max ∑σ
j∈J cp
j
©  J.  Christopher  Beck 7
PERT  Practice
n Draw  precedence   Job paj pmj pbj Predecessors
1 2 4 12 -­
graph 2 10 15 20 1
n Find  μj  ,  δj   3 6 8 22 1
(see  P.  4.3) 4 8 16 18 1
5 2 10 18 2,3,4
n Find  Critical  Path(s) 6 8 12 24 2
n Estimate  expected   7 2 5 8 5
8 3 4 11 5
value  and  variance  
9 4 8 24 6,7
of  Cmax 10 1 5 9 8
©  J.  Christopher  Beck  2015 8
More  PERT  Practice
Example  4.3.1
Jobs 1 2 3 4 5 6 7 8 9 10 11 12 13 14
paj 4 4 8 10 6 12 4 5 10 7 6 6 7 2
pmj 5 6 8 11 7 12 11 6 10 8 7 8 7 5
pbj 6 8 14 18 8 12 12 7 10 15 8 10 7 8

Hint:  same 2 4 7 Find  expected  


graph  as   1 10 makespan  and
4.2.3 6 9 12 variance
3 14
11
©  J.  Christopher  Beck  2015 5 8 13 9
PERT  Problems
n More  than  one  CP?
n non-­CP  with  high  variance?
n expected  makespan  must  be
larger  than  single  CP  estimate  (why?)
n Assumption  of  normal
distribution

©  J.  Christopher  Beck 10

Vous aimerez peut-être aussi