A controversy has developed over the years
between the proponents of PERT and CPM. The commonly debated questions are:
Which of these methods originated first? Which is the better one? Are both of
them practically the same? To understand the various distinguishing features of
the two and their spheres of application, let us first discuss their original
differences and then the subsequent developments. All controversial issues,
which are only of an academic interest with no practical significance have been
avoided.
Original Differences
CPM was developed for
planning, scheduling and control of civil works, while PERT originated in
response to the complexities of the uncertainity in research and development
projects for controlling their multifarious time schedules. Originally, thus,
their fields of application were quite different.
In network modelling, CPM laid emphasis on
breaking the projects into various works or activities. In PERT, the project
breakdown was in terms of milestones which were planned to occur during its
execution. Therefore, CPM was activity-oriented whereas PERT was
event-oriented.
Originally, the application of CPM was
confined to construction works where the activities were familiar and their
duration could be easily estimated from the one-time estimate. Since PERT was
designed to cope with uncertainties, it used the three-time estimate.
In CPM, activity durations where related
to costs. This provided a means of assessment of different activity durations
with varying costs and made crashing of activities possible. PERT dealt with
events and their probable time of occurrence. This enabled adoption of
probabilistic approach in time scheduling.
The CPM schedule enabled optimization of
resources as the activity durations were defined in terms of resources
employed. This was not possible with PERT.
CPM, which used the one-time estimate, was
simpler to follow, while PERT required a statistician to interpret the results.
Later developments
As the various drawbacks of PERT gradually
became apparent, further studies were initiated by the US Defence Department.
Notable among these were the PERT/TIME and PERT/COST. In the later, the concept
of single-time estimate of activities could be equally employed.
In the CPM field, the concept of three-time
estimate was introduced into networks which were primarily designed for
controlling the time factor rather than the resources. In recent years, further
developments in PERT and CPM have made CPM appear more like PERT, and the
subject of network analysis has come to be known as PERT/CPM. To distinguish
these two main network techniques, the event-oriented networks using three-time
estimate for activity durations having uncertainties can be termed as a PERT
network, while the other which is activities-oriented and uses one-time
estimate may be referred to as the CPM.