Scribd
Upload
697 views

Monte Carlo Simulation

Monte Carlo simulation is an alternative to PERT for evaluating risks of not achieving project deadlines. It involves calculating activity completion times for a project network many times by randomly selecting estimated activity times. This provides a representative selection of possible completion times. The critical chain approach schedules each activity as late as possible while still meeting the target date. It inserts project buffers after the critical chain and feeding buffers after subsidiary chains to absorb delays. When executed, no chain should start early and tasks must be finished as soon as predecessors are complete. Buffers are divided into green, amber and red zones to monitor schedule risk.

Uploaded by

VinayKumarSingh
Copyright
© © All Rights Reserved
Available Formats
Download as PPT or read online on Scribd
697 views

Monte Carlo Simulation

Monte Carlo simulation is an alternative to PERT for evaluating risks of not achieving project deadlines. It involves calculating activity completion times for a project network many times by randomly selecting estimated activity times. This provides a representative selection of possible completion times. The critical chain approach schedules each activity as late as possible while still meeting the target date. It inserts project buffers after the critical chain and feeding buffers after subsidiary chains to absorb delays. When executed, no chain should start early and tasks must be finished as soon as predecessors are complete. Buffers are divided into green, amber and red zones to monitor schedule risk.

Uploaded by

VinayKumarSingh
Copyright
© © All Rights Reserved
Available Formats
Download as PPT or read online on Scribd
You are on page 1/ 18

MONTE CARLO

SIMULATION
• Alternative to PERT technique.
• Use to evaluate the risks of not achieving
deadlines.
• Basis of this technique involves
– Calculating activity completion times for a
project network a large no: of times.
– Selecting estimated activity times randomly
from a set of estimates for each activity.
25

20

15 Activity start
date
10 Activity end
date
5

0
56 59 62 65 68 71 74 77 80
• For calculation, we may have to carry out
the forward pass through the network
many times before obtaining a
representative selection of possible
completion times.
• Many packages for carrying out Monte
Carlo Simulation
• Exchange data with project-scheduling applications
• Interface to standard spreadsheet software
Critical chain concepts
• When the project is actually being executed
,the project mgr will be forced to focus on
the activities where the actual durations
exceed the target.
• Activities which are completed before the
target data are likely to be overlooked.
Weeks that activities were early or late

35
Percentage of activities

30

25

20

15

10

0
-6w -5w -4w -3w -2w -1w 0w 1w 2w 3w 4w 5w 6w

weeks

Percentage of activities early of stage


• Findings of Michiel van Genuchten, the reasons
for delays in the completion of software
development task.
• About 30% activities were finished on time, while
9% were a week early & 17% were a week late.
• The jump of 21% points b/w being a week early & being
on time is compatible with the ‘Parkinson’s Law’s
principle’
• ‘work expands to fill the time available’.
• Reason for delay was the time that had to be spent on
non-project work.
• One approach which attempts to solve
some of these problems is the application
of the Critical chain concept.
• To demonstrate the principles of this
approach we use Gantt chart
• A ‘traditional approach’ were adopted &
• Pessimistic durations were used to supply ‘comfort
times’ for each activity.
3rd Quarter 4th Quarter 1st Quarter 2nd Quarter 3rd Quarter
ID i
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

1 A. Hardware selection 8 wks

2 B. Software design 5 wks

3 C. Install hardware 3 wks

4 D. Code/test software 5 wks

5 E. File take on 4 wks

6 F. Write user manuals 15 wks

7 G. User training 4 wks

H. Install/test system 2.5 wks


8

Gantt chart – ‘traditional’ planning approach.


– General steps in the Critical chain approach
• Target date given to the developer is one where it
is estimated that there is a 50% chance of
success-this approximates to the expected time
identified in the PERT risk method.
• Working backwards from the target completion
date, each activity is scheduled to start as late as
possible. This should reduce the chance of staff
being pulled off the project on to other work. It also
make every activity ‘Critical'. If it is late the whole
project is late.
3rd Quarter 4th Quarter 1st Quarter 2nd Quarter 3rd Quarter
ID i
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

1 A. Hardware selection 6 wks

2 B. Software design 4 wks

3 C. Install hardware 3 wks

4 D. Code/test software 4 wks

5 E. File take on 3 wks

6 F. Write user manuals 10 wks

7 G. User training 3 wks

H. Install/test system 2 wks


8

Gantt chart – plan using ‘expected times’ & starting all activities as late as possible
– In the Gantt chart, all the most likely durations
have been used & every activity is scheduled
to start as late as possible as long as the
project is completed by 1st March.
• To cope with activity overruns, a project buffer is
inserted at the end of the project before the target
completion date. This buffer is the equivalent of
50% of the length of the ‘comfort zone’ that has
been removed from the critical chain.
– Longest chain of activities in the project, taking account
of task & resource dependencies.
Different from critical path, only take account of
task dependencies.
Resource dependency-one activity has to wait for
a resource which is being used by another
activity to become available.
If an activity is late it will push the project
completion date further into the project buffer.
• Buffers are also inserted into the project
schedule where a subsidiary chain of activities
feeds into the critical chain.
– Feeding buffer -50% of the length of the
‘comfort zone’ removed from the subsidiary or
feeding chain.
3rd Quarter 4th Quarter 1st Quarter 2nd Quarter 3rd Quarter
ID i
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

1 A. Hardware selection 6 wks

2 B. Software design 4 wks


3 C. Install hardware 3 wks
4 D. Code/test software 4 wks

5 E. File take on 3 wks


6 Feeding buffer 1 wks
7 F. Write user manuals 10 wks
8 G. User training 3 wks
9 H. Install/test system 2 wks
10 Feeding buffer 1.25 wks

11 Project buffer 3 wks

Gantt chart – feeding buffers added


• Critical chain same as critical path, i.e. activity
F &G which have comfort zones of 5 & 1
weeks, making a total of 6 weeks.
Then project buffer is 3 weeks.
Subsidiary chain feeds into the critical chain –
activity H links into the project buffer & activity
E links into G which is part of critical chain.
For the first buffer the duration would be 50% of
the saved the comfort zone of A,C and H that
is (2+0+.5)/2=1.25 weeks. In the second case
it would be 50% of the comfort zones of
activities Band E that is 1 week.
• When the project is executed the following
principles are followed.
• No chain of tasks should be started
earlier than scheduled, but once it has
been started it should be finished as soon
as possible-this invokes the relay race
principle, where developer should be
ready to start their tasks as soon as the
previous,dependent,tasks are completed.
• Buffers are divided into three
zones:green,amber,and red each of an even
(33%)size
1. green, where no action is required if the project
completion date creeps into this zones
2. Amber, where an action plan is formulated if the
project completion dates moves into this zone
3. Red, where the action plan above is executed if the
project penetrates this zone.

You might also like