Lecture 07
Lecture 07
nd
Topics
Chapters
# of lectures
1. Project Management
2. Inventory
16
1.5
3. Decision Analysis
12
1.5
4. Queuing
13
1
(to p2)
2
(to p3)
Project Management
Questions:
1. Why do we need to study Project
(to p4)
Management?
2. How does a project management
(to p5)
technique work?
3
Objective
The main purpose is to govern the
operations of a project such that all
activities involved are well administrated
and that we can also control its
completion time
4
(to p3)
(to p29)
2.
Special case!
Event
Processing
Time
Precedent
constraints
A
B
C
20
30
10
-A
B
(to p9)
(to p4)
(to p8)
(to p6)
Example
Path
Event
Proc
Time
Pred
Const
1-2
2-3
3-4
A
B
C
20
30
10
-A
B
A
1
2
20
B
3
30
4
10
(to p7)
Special case!
Event Processing Precedent
When two or events
Time
constraints
taken places in the
same time interval
(known an concurrent A
3
-events)
B
5
A
Consider the following C
7
A
example!
(to p10)
Case 1
A
1
C 7
Wrong!
Rule2: no node can have
two outcomes and end
with the same note
Solution
(to p11)
10
3
C
5
4
Solution 2:
What one
is better?
Dummy = 0
3
B
A
1
Dummy 2 = 0
B
C
Solution 3:
Dummy 1=0
3
C
Dummy = 0
(to p6)
11
12
Figure8.3
Expanded network
for building a house
showing
concurrent
activities.
Table8.1
Possible Paths to
complete the
House-Building
Network
(to p14)
13
This is the
Solution!
path C: 1 2 4 6 7, 3 + 1 + 3 + 1 = 8 months
path D: 1 2 4 5 6 7, 3 + 1 + 1 + 1 + 1 = 7 months
The critical path is the longest path through the network; the minimum time the network can be completed.
Figure8.5
Alternative paths in the
network
(to p12)
14
How it works?
(to p16)
15
Branch
ES EF LS LF
with
EF1=0
16
(to p4)
Compute ES and EF
Note:
When computing these values, the pattern
is like moving zic-zac format by firstly
computer ES12 and then adding it to EF12
and move to next branch by copying the
max values of the branch 1-2 to say, 2-3
We compute them from top to bottom!
(to p18)
Their relationship :
(to p19)
(to p22)
Example 1:
17
Then
t12
EF1 = 0
ES12 = max (EF1)
=0
18
Branches
ESij = max(EFi)
EFij=ESij+tij
1-2
2-3
2-4
3-4
4-5
4-6
5-6
6-7
ES12= max(EF1)=
EF12=ES12+t12=
ES23=max(EF2)=
EF23=ES23+t23=
ES24=max(EF2)=
EF24=
ES34=max(EF3)=
EF34=
ES45=max(EF4)=
EF45=
ES46=max(EF4)=
EF46=
ES56=max(EF5)=
EF56=
ES67=max(EF6)=
The overall computation is shown in next slide
EF67=
(to p20)
19
Complete solution
(note:you can compute these values and show in the network diagram as well)
(to p21)
=max(5+3,5+1)=8
20
(to p4)
Figure8.6
Earliest activity start and finish times
(to p20)
21
Compute LS and LF
Note: We compute these values from the
bottom to top, with assigning:
LSij = LFi -tij
with
the end of LFij = EFij
(to p23)
22
Branches
LSij = LFij-tij
LFij=min(LSj)
1-2
2-3
2-4
3-4
4-5
4-6
5-6
6-7
LS12 = Li12-t12 =
LS23 = LF23-t23 =
LS24 = LF24-t24 =
LS34 = LF34-t34 =
LF12=min(LS2)=
LS45 = LF45-t45 =
LS46 = LF46-i46 =
LS56 = LF56-t56 =
LS67 = LF67-t67 =
LF45=min(LS5)=
LF23=min(LS3)=
LF24=min(LS4)=
LF34=min(LS4)=
LF46=min(LS6)=
LF56=min(LS6)=
LF67=min(LS7)=
23
(to p24)
LS is the latest time an activity can start without delaying critical path time. LSij = LFij - tij
- LF is the latest finish time
Min(node3-t23,node4-t24)
=Min(5-2,5-1)=Min(3,4)=3
Min(node 6-t46,node5-t45)
=Min(8-3,7-1)
=Min(5,6)=5
Min(node 7-t67)
=Min(9-1)=8
Start with the end node first
Same as EF67
from the previous slide
24
(to p25)
(to p22)
Figure8.7
Latest activity start and finish times
(to p24)
25
(to p27)
2.
by showing branches
(to p28)
26
(to p16)
Figure8.9
Activity Slack
Slack time exists for those activities not on the critical path for which the earliest and latest star
times are not equal.
Shared slack is slack available for a sequence of activities.
Figure8.8
Earliest activity start and finish times
28
(to p26)
Sensitivity Analysis
Today, we only consider one case
Probabilistic Activity Times
Refer to activity time estimates usually can
not be made with certainty
(to p30)
PERT
In PERT, three different time estimations are
applied:
most likely time (m),
the optimistic time (a) , and
the pessimistic time (b).
How do we make use of these three values?
(to p31)
30
variance:
a 4m b
6
b-a
v
(to p32)
31
PERT
We simply apply t values in CPM and determine the
values of:
ES
EF
LS
LF
S
Example.
(to p33)
32
Example
Result implication
(to p39)
Applications
(to p38)
33
PERT Example
Step 1: computer t and v values
Step 2: determine the CPM
Step 3: determine v value
(to p35)
(to p36)
(to p37)
(to p33)
34
a 4m b
6
b-a
v
Figure8.11
Network with mean activity times and variances
Table8.3
Activity Time Estimates for
Figure 8.10
35
(to p34)
Figure8.12
Earliest and latest activity times
Table8.4
Activity Earliest and
Latest Times and Slack
36
(to p34)
13
35
57
7 9
Variance
total
1
1/9
16/9
4
62/9
37
(to p34)
Figure8.13
Normal distribution of network duration
Critical value
38
(to p33)
Consider when
x = 30
(to p40)
x = 22
(to p41)
Tutorial Assignment
(to p42)
39
40
(to p39)
Figure8.15
Probability the network
will be completed in 22
weeks or less
41
(to p39)
Tutorial Assignment
Try to use QM to solve CPM/PERT
(to p43)
problems (see slide 19)
Exercises (Chapter 8)
Old: 8, 10, 17
New: 4, 6, 11
42
Exhibit 8.1
43
(to p16)
Slack time exists for those activities not on the critical path for which the earliest and latest star
times are not equal.
Shared slack is slack available for a sequence of activities.
Figure8.8
Earliest activity start and finish times
44
Figure8.9
Activity Slack
45