You are on page 1of 31

Lecture 4:  

Time/Cost  Trade-­off  in  


Project  Planning

©  J.  Christopher  Beck  2015 1


Outline
n Project  Planning  with  
Time/Cost  Tradeoffs
n Linear  and  non-­linear

©  J.  Christopher  Beck  2015 2


Readings
n P  Ch  4.4-­4.5
n Slides  borrowed
from  Twente  &  
Iowa
n See  Pinedo  CD

©  J.  Christopher  Beck  2015 3


Time/Cost
Trade-­Offs
n What  if  you  could  
spend  money  to  
reduce  the  job  
duration
n More  money   shorter  
processing  time
n e.g.,  run  machine  at  
higher  speed

©  J.  Christopher  Beck  2015 4


Linear  Costs
Cost  to  reduce  job  j  by p min ≤ p ≤ p max
Money j j j
by  1  time  unit
c aj
c aj − c bj
Marginal c j = max min
cost p j − p j
c bj

Processing
min max time
p j p j

©  J.  Christopher  Beck  2015 5


Problem
n Spend  money  to  reduce  processing  
times  so  as  to  minimize:
n
( (
c0Cmax + ∑ c bj + c j p max
j − pj ))
j =1

“Overhead”  cost Marginal  cost  per  activity


(cost  per  time  unit,  
independent  of  activities:
worker  salaries  or
electricity  for  lights)

©  J.  Christopher  Beck  2015 6


Solution  Methods
n Objective:  minimum  cost  of  project
n Linear  programming  formulation
n Optimal  schedules
n Costs  must  be  linear
n Time/Cost  Trade-­off  Heuristic
n Good  schedules
n Works  also  for  non-­linear  costs

©  J.  Christopher  Beck  2015 7


LP  Formulation
Decision  variables?

How  do  the  xj’s


get  a  value?

3
©  J.  Christopher  Beck  2015 4 8
Cost  Function
n How  do  we  go  from:
n
(
c0Cmax + ∑ c + c j p b
j ( max
j − pj ))
j =1

n to:
n

c 0Cmax − ∑ c j p j
j =1

©  J.  Christopher  Beck  2015 9


Solution  Methods
n Objective:  minimum  cost  of  project
n Linear  programming  formulation
n Optimal  schedules
n Costs  must  be  linear
n Time/Cost  Trade-­off  Heuristic
n Good  schedules
n Works  also  for  non-­linear  costs

©  J.  Christopher  Beck  2015 10


Sources,  Sinks,  &  Cuts

Cut set

Sink node

Source (dummy) node


Minimal cut set
©  J.  Christopher  Beck  2015 11
Cuts  Sets  &  Minimal  Cut  Sets
V  is  a  set  of  nodes
n Given  a  graph  G(V,E) E  is  set  of  edges  (arcs)
n A  cut  set is  a  set  of  nodes S ⊂ V
such  that  in  the  graph  formed  by  
removing  the  nodes  in  S  from  G  there  is  
no  path  from  the  source  to  the  sink
n A  minimal  cut  set is  i)  a  cut  set  such  
that  ii)  none of  its  proper  subsets  is  a  
cut  set
©  J.  Christopher  Beck  2015 12
Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 13


Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 14


Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 15


Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 16


Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 17


Is  This  a  Cut  Set?
Is  This  a  Minimal  Cut  Set?

2 4 7
10
1
6 9 12 14

3 11 13

©  J.  Christopher  Beck  2015 18


Time/Cost  Tradeoff  Heuristic
n Idea:  we  want  to  repeatedly  find  a  set  
of  activities  which  we  can  make  shorter
n Must  result  in  shorter  makespan Why?
n Must  result  in  better  solution  (i.e.,  lower  
cost)
n Once  we  make  them  shorter,  look  for  
another  set

©  J.  Christopher  Beck  2015 19


Time/Cost  Trade-­off  Heuristic
Step 1:
Set all processing times at their maximum
max
pj = p j

Determine all critical paths


Construct the graph Gcp of critical paths

©  J.  Christopher  Beck  2015 20


Time/Cost  Trade-­off  Heuristic
Step 2:
Determine all minimal cut sets in Gcp
Consider those sets, C, where all processing times are
larger than their minimum
min Why?
pj > p j ,∀j ∈ C

If no such set STOP; otherwise continue to Step 3


©  J.  Christopher  Beck  2015 21
Time/Cost  Trade-­Off  Heuristic
Step 3:
For each minimal cut set:
Compute the cost of reducing all processing times by one
time unit.
Take the minimal cut set with the lowest cost
If this is less than the overhead per time unit go on to Step
4; otherwise STOP

©  J.  Christopher  Beck  2015 22


Time/Cost  Trade-­Off  Heuristic
Step 4:
Reduce all processing times in the minimal cut set by one

time unit

Determine the new set of critical paths

Revise graph Gcp and go back to Step 2

©  J.  Christopher  Beck  2015 23


Example  4.4.2 Error  in  text

Jobs 1 2 3 4 5 6 7 8 9 10 11 12 13 14

pmaxj 5 6 9 12 7 12 10 6 10 9 7 8 7 5

pminj 3 5 7 9 5 9 8 3 7 6 4 5 5 4

caj 20 25 20 15 30 40 35 25 30 20 25 35 20 10

cj 7 2 4 3 4 3 4 4 4 5 2 2 4 8

©  J.  Christopher  Beck  2015 24


Warning:  Example  4.4.2

n Unfortunately,  the  text  is  completely  


wrong  for  this  example!

©  J.  Christopher  Beck  2015 25


1.  Step  1:  Maximum  Processing  
Times,  Find  Gcp

2 4 7
10
1
6 9 12 14

3 11 13

5 8

©  J.  Christopher  Beck  2015 26


1.  Step  1:  Maximum  Processing  
Times,  Find  Gcp
Cmax = 56
2 4 7
10
1
6 9 12 14

3 11 13
Cost  =  overhead  +  job  costs 5 8
=  co *  Cmax +  Σcbj
=  6  *  56  +  225
=  561  

©  J.  Christopher  Beck  2015 27


1.  Step  2  &  3:  Min.  Cut  Sets  in  
Gcp &  Lowest  Cost
c1=7
c6=3 c9=4 c12=2 c14=8
1
6 9 12 14

3 11
c3=4 c11=2

Cut sets: {1},{3},{6},{9},


{11},{12},{14}. Minimum cut
Arbitrarily choose 12 sets with lowest cost
Reduce p12 from 8 to 7

©  J.  Christopher  Beck  2015 28


Next  Iteration?

©  J.  Christopher  Beck  2015 29


Example  4.4.2
n If  you  are  lost,  you  have  work  to  do.
n Remember  the  book  is  wrong.  You  
should  use  the  slides.

©  J.  Christopher  Beck  2015 30


Can  Also  Have
Non-­linear  Costs
n Arbitrary  function   cj(pj)  →  cost  of  
setting  job  j  to  processing  time  pj
n LP  doesn’t  work!
n See  Section  4.5
n A  question  I  like:
n Given  processing  times  and  cj(pj),  which  
algorithm  would  you  use  (heuristic  or  LP)?

©  J.  Christopher  Beck  2015 31

You might also like