Introduction
PERT is a network analysis technique used to estimate project duration when there is a high degree of uncertainty about the individual activity duration estimates. PERT uses probabilistic time estimates based on using optimistic, most likely, and pessimistic estimates of activity durations
The Program Evaluation and Review Technique (PERT) is a widely used method for planning and coordinating large-scale projects. As Harold Kerzner explained in his book Project Management, “PERT is basically a management planning and control tool. It can be considered as a road map for a particular program or project in which all of the major elements (events) have been completely identified, together with their corresponding interrelations’. PERT charts are often constructed from back to front because, for many projects, the end date is fixed and the contractor has front-end flexibility.” A basic element of PERT-style planning is to identify critical activities on which others depend. The technique is often referred to as PERT/CPM, the CPM standing for “critical path method.”
PERT was developed during the 1950s through the efforts of the U.S. Navy and some of its contractors working on the Polaris missile project. Concerned about the growing nuclear arsenal of the Soviet Union, the U.S. government wanted to complete the Polaris project as quickly as possible. The Navy used PERT to coordinate the efforts of some 3,000 contractors involved with the project. Experts credited PERT with shortening the project duration by two years. Since then, all government contractors have been required to use PERT or a similar project analysis technique for all major government contracts.
Network Diagrams
The chief feature of PERT analysis is a network diagram that provides a visual depiction of the major project activities and the sequence in which they must be completed. Activities are defined as distinct steps toward completion of the project that consume either time or resources. The network diagram consists of arrows and nodes and can be organized using one of two different conventions. The arrows represent activities in the activity-on-arrow convention, while the nodes represent activities in the activity-on-node convention. For each activity, managers provide an estimate of the time required to complete it.
The sequence of activities leading from the starting point of the diagram to the finishing point of the diagram is called a path. The amount of time required to complete the work involved in any path can be figured by adding up the estimated times of all activities along that path. The path with the longest total time is then called the “critical path,” hence the term CPM. The critical path is the most important part of the diagram for managers: it determines the completion date of the project. Delays in completing activities along the critical path necessitate an extension of the final deadline for the project. If a manager hopes to shorten the time required to complete the project, he or she must focus on finding ways to reduce the time involved in activities along the critical path.
The time estimates managers provide for the various activities comprising a project involve different degrees of certainty. When time estimates can be made with a high degree of certainty, they are called deterministic estimates. When they are subject to variation, they are called probabilistic estimates. In using the probabilistic approach, managers provide three estimates for each activity: an optimistic or best case estimate; a pessimistic or worst case estimate; and the most likely estimate. Statistical methods can be used to describe the extent of variability in these estimates, and thus the degree of uncertainty in the time provided for each activity. Computing the standard deviation of each path provides a probabilistic estimate of the time required to complete the overall project.
PERT Analysis
Managers can obtain a great deal of information by analyzing network diagrams of projects. For example, network diagrams show the sequence of activities involved in a project. From this sequence, managers can determine which activities must take place before others can begin, and which can occur independently of one another. Managers can also gain valuable insight by examining paths other than the critical path. Since these paths require less time to complete, they can often accommodate slippage without affecting the project completion time. The difference between the length of a given path and the length of the critical path is known as slack. Knowing where slack is located helps managers to allocate scarce resources and direct their efforts to control activities.
For complex problems involving hundreds of activities, computers are used to create and analyze the project networks. The project information input into the computer includes the earliest start time for each activity, earliest finish time for each activity, latest start time for each activity, and latest finish time for each activity without delaying the project completion. From these values, a computer algorithm can determine the expected project duration and the activities located on the critical path. Managers can use this information to determine where project time can be shortened by injecting additional resources, like workers or equipment. Needless to say, the solution of the algorithm is easy for the computer, but the resulting information will only be as good as the estimates originally made. Thus PERT depends on good estimates and sometimes inspired guesses.
PERT offers a number of advantages to managers. For example, it forces them to organize and quantify project information and provides them with a graphic display of the project. It also helps them to identify which activities are critical to the project completion time and should be watched closely, and which activities involve slack time and can be delayed without affecting the project completion time. The chief disadvantages of PERT lie in the nature of reality. Complex systems and plans, with many suppliers and channels of supply involved, sometimes make it difficult to predict precisely what will happen. The technique works best in well-understood engineering projects where sufficient experience exists to predict tasks accurately in advance.
