Allocating Scarce Resources to Projects

When we leveled resources in the case of the overworked scriptwriter, MSP simply used the available activity slack to reschedule WBS item 3.2 (see Figure 6-13). The project completion date was not altered because the WBS 3.2’s slack was large enough to swal­low the added time. Often, that is not possible and the software needs instruction about what priority to use when allocating scarce resources to several tasks—which tasks should get the resources immediately and go first, and which may be delayed. In order to select a suitable method for assigning priorities, we need to understand how the problem is solved.

Most solutions start with the PERT/CPM schedule. Given this initial solution, each activity is examined period by period and resource by resource. If the demand for a resource exceeds its supply, the software considers the tasks one by one and assigns resources to the tasks according to some priority rule chosen by the PM. Tasks that receive resources under this rule proceed as originally scheduled. Tasks that do not get resources are delayed until ongoing tasks are completed and the required resources are freed-up for use. If this increases the project duration, the change will be visible on the project’s Gantt chart or network as it was when the videotape project client was not available while on vacation. (See Figures 6-7, 6-8, and 6-13.) No matter what priority rule is used, the project’s technology always takes precedence.

1. Some Comments about Constrained Resources

Every time a project falls behind schedule, the PM is apt to plead for more resources. In spite of the PM’s complaints about the scarcity of resources, serious cases of resource scar­city rarely apply to resources in general, but rather to one or two very specific resources. We call such resources “Walts.” The term was derived from the name of an individual, Walter A., who is employed by a large insurance company. Walt is a specialist in the rules and laws affecting insurance policies for certain types of casualty losses in the firm’s com­mercial lines of business. He has an excellent analytical mind and many years of experi­ence. His knowledge is required when designing new policies in this area of risk. The firm has only one Walt, and while the firm is training others, such training takes years. Walt is a true scarce resource. Projects requiring Walt’s input are scheduled around his availability.

There are many other examples. Military combat missions may be scheduled around the availability of attack aircraft. Construction projects may be scheduled around the availability of a large crane. A Broadway opening may be scheduled around the availabil­ity of a star actress. The key problem to be solved is deciding which activities get the scarce resource and in what order.

2. Some Priority Rules

There are many possible rules for assigning preference to some activities over others when allocating scarce resources. Most popular project management software packages have a limited number of rules that can be automatically applied to level overallocated resources so many of the priority rules for assigning scarce resources to activities may have to be applied manually. Fortunately, as we will see, this is not as difficult as it might seem. Several of the most commonly used rules are as follows:

As soon as possible—This is the standard rule in scheduling. Activities are sched­uled to start on their earliest start times (ES), and resources are made available with that in mind.

As late as possible—With this rule, resources are made available so that activities start on their latest start times (LS) whenever possible without increasing the project’s duration. This may seem irrational, but it preserves the firm’s resources and delays cash outflows as long as possible. This rule is also compatible with Eliyahu Goldratt’s conten­tion that the “student syndrome” leads workers to delay starting an activity until the last possible moment (Goldratt, 1997, Ch. 13).

Shortest task duration first—Always consistent with technological precedences, shorter tasks are given priority over longer tasks. This rule maximizes the number of tasks that can be completed by a system in a given time period.

Minimum slack first—Tasks are supplied with resources in inverse order of their slacks. This rule usually minimizes the number of late activities.

Most critical followers—The number of successors on the critical path(s) for each activity is counted. Activities with a higher number of critical successors take prece­dence. The rationale here is that such activities would cause the greatest damage to the desired project schedule if they are late.

Most successors—The same as the previous rule except that all successors are counted. This rule has the same rationale as the preceding rule.

Most resources first—With this rule, the greater the use of a specific resource on a task, the higher the task’s priority for the resource. This rule is based on the assumption that more important activities have a greater demand for scarce resources.

In addition to these rules, there are many others. For example, it may be company policy to put favored customers’ projects at the head of the resource line—or to reserve special resources for such clients by withholding them from the available supply. The same type of favoritism is sometimes shown to specific projects of high value to the par­ent firm. (Some firms show favoritism to specific high-value activities, but this rule makes little sense because all activities of an individual project must be completed to finish the project.) Application of a value measure for allocating scarce resources across several projects is both rational and common.

There are many other priority rules that might be used, but most project manage­ment software packages recognize only a few; however, assigning scarce resources manu­ally is not difficult. From our earlier example, recall Figure 6-9, in which all tasks requiring the scriptwriter were listed alone on a Gantt chart showing the task duration, scheduled start, scheduled finish, activity slack, and other needed information. The scheduled start dates would allow us to apply either the early or late start rules. Information on activity slack allows us to use the minimum slack rule. Task duration is the necessary input to apply the Shortest Task Duration rule. The project network, Gantt chart, and the task list all allow a simple count of followers, critical or not. MSP and most other project management software allow verbal or numeric priorities to be assigned easily. This allows the use of any priority system required.

Considerable research has been done on these rules[1] and the minimum slack rule is usually best or second best. It rarely performs poorly. If a high-slack task is not given resources in one period, its slack is automatically decreased and in the next period it has a better chance of receiving resources. The resource allocation is repeated periodically (hourly, daily, weekly, or monthly, etc.), depending on the time frame of the project’s activities. If a task becomes critical, that is, all the slack is used up before the activity receives resources, the project will be delayed. We will consider borrowing resources from ongoing tasks when we discuss the allocation of scarce resources among sev­eral projects.

Source: Meredith Jack R., Mantel Jr. Samuel J., Shafer Scott M., Sutton Margaret M. (2017), Project Management in Practice, John Wiley & Sons, Inc. 3th Edition.