Program Evaluation Review Technique (PERT)
Program
Evaluation Review Technique (PERT) charts show each task in a project as a
node. Dependencies between tasks (e.g. where one task requires another one to
be completed before it can start) are clearly shown by interconnections between
the task nodes. PERT charts also show timing information for each task. PERT
charts are similar to the critical path method (CPM) which identifies the
longest path through the project, and therefore the minimum time for the
project to be completed. In the example below, the critical path is shown by
the red arrows.
Project
managers can use PERT charts to:
- set a realistic timetable for project completion
- make sure focus is maintained on the most critical tasks for the critical path – since the path leads to the minimum time the project requires, any delays to these tasks will result in a delay to the overall project
- identify tasks that need to be shortened if the overall project time needs to be reduced
- identify tasks that can be carried out simultaneously
- identify slack time where certain tasks are not as time-critical to the overall deadline.
Whereas CPM uses a single estimate for task duration, the PERT
method requires the project manager to supply three estimates to make the
estimates as accurate as possible. The three estimates are:
· Most Likely Time (M) – the project manager should have a high degree of certainty that the task will be completed within this time
· Pessimistic Time (P) – the project manager should be almost certain (e.g. 99%) that the task will be completed within this time
The ‘expected time’ 𝑇𝐸 is then calculated as 𝑇𝐸 = 𝑂+4𝑀+𝑃/6
The following
table shows the tasks, dependencies, and estimated times a project manager
might input to a PERT chart for a software development project and then calculated Expected Times:
Task Identifier
|
Task Description
|
Predecessor Task(s)
|
Optimistic Time (O)
|
Most Likely Time (M)
|
Pessimistic Time (P)
|
Expected Time (TE)
|
A
|
Establish project
|
4
|
5
|
12
|
6
|
|
B
|
Establish customer requirements
|
A
|
2
|
3
|
4
|
3
|
C
|
Produce software specification
documents
|
B
|
6
|
8
|
22
|
10
|
D
|
Write test plans
|
C
|
4
|
6
|
8
|
6
|
E
|
Write code
|
C
|
3
|
4
|
5
|
4
|
F
|
Developer testing
|
E
|
2
|
4
|
6
|
4
|
G
|
System testing
|
D, F
|
2
|
3
|
4
|
3
|
H
|
Write manuals
|
C
|
5
|
7
|
15
|
8
|
Network Diagram:
The network diagram is based on the tasks and their dependencies (predecessor tasks).Task A has no predecessor, and therefore starts the project on the left. Task B has only task A as a predecessor, and is therefore the next task. The diagram shows that tasks D, E and H (writing test plans, code and manuals) all have task C (specification) as a predecessor, and can therefore be carried out simultaneously.
Identify the critical path
The critical path is determined by using the estimated times to work
out the earliest start (ES) and finish (EF) times (forward pass), and latest
start (LS) and finish (LF) times (backward pass), and identifying the tasks
where ES and LS are equal.
Step1: Forward
pass
This is done by working from left to right across the chart. The ES
of task A is week 0, and the EF is 6 since this is the ES plus the duration
(estimated time). The earliest that task B can begin is week 6 (since task A
must be complete) so its ES is 6, and EF is 6 + 3 = 9 . Note that task G has an
ES of 27 because it is dependent on both task D and task F being complete
first, and the earliest that both tasks will be completed is week 27.
Step 2:
Backward pass
Working from
right to left, the Latest Finish Time (LF) and Latest Start Time (LS) can be
filled in. The final time is 33 weeks, so the LF of task H must be 33. Because
this task has a duration of eight weeks it means it has an LS of 33 – 8 = week
25.
Note that for some tasks, such as Task C, there are different
possibilities. To calculate the LF of Task C we need to look at the latest
start times for Tasks D, E and H which are dependent on it. The latest time
task C can finish (LF) must be 19 weeks because Task E has to start at this
point.
Step 3: Critical path and slack
The critical path is identified by finding the nodes where ES = LS.
These are nodes for which there is no ‘slack’ or spare time for a task in the
project.
The slack is calculated as LS – ES, e.g. Task H has 6 weeks’ slack,
and could be scheduled to start sometime between week 19 and week 25 depending
on factors such as the resources available to a task on the critical path.