This preview shows page 1. Sign up to view the full content.
Unformatted text preview: f the schedule is the best policy, because, even in the absence of variances, tasks can complete late for various reasons, and this tendency alone generally causes schedule growth. Many programs reschedule monthly, but shorter or longer periods may be more appropriate. As a general rule of thumb, the frequency of rescheduling should increase if schedule growth is increasing and can be reduced if the total duration is contracting. Just as different parts of an organization need to see different views of the performance and productivity reports, they may also only want to see the part of the scheduling network Page 188 for which they have responsibility. Scheduling systems like Microsoft Project have the capability to produce subsets of the schedule, sometimes referred to as subnets. Often, these crafts or departments will want to do a critical path analysis of a subnet, but we will not discuss all of the possible variations here. The point to be made is that managers on the project need to be aware of the part of the project schedule that applies to them and of the intersection of the critical path with this part of the schedule. 7.2.3— Resource Scheduling So far, all of our discussion of scheduling has centered on simple scheduling. We have seen that a simple scheduling system usually computes two sets of dates for each task. For tasks on the critical path, these two sets of dates are the same. For all those tasks that are not on the critical path, it is possible to schedule them between these two sets of dates. By scheduling these noncritical tasks later than their early start dates, it is possible to reduce the demand for resources for all tasks that need these resources. For example, suppose we have two tasks t1 and t2 such that t1 is on the critical path but t 2 is not. Further suppose that both t 1 and t2 need resource R during their entire duration, but there is only enough of R to fulfill the needs of one of the tasks. If by chance t1 and t 2 have the same early start date, it would be possible, but unwise, to start t1 and t2 at the same time, because the demand placed on R by t2 would take away from its ability to serve t 1, possibly causing t 1 to finish late. Since t1 is on the critical path, finishing late will cause the total duration of the project to slip. Consequently, we see that resources should always be allocated to critical tasks before they are allocated to noncritical tasks. Originally, the process of scheduling the noncritical tasks to minimize the demand on all resources was called resource leveling . Somewhat later, it was realized that guessing at the durations of tasks also produced less than optimal schedules. If Page 189 the scheduling system knew how much of each resource it needed to complete each task, it might be possible for it to compute an optimal schedule consisting of the best duration set possible for all the tasks that both minimized the schedule and the consumption of resources. Unfortunately, in general, this task scheduling problem has been shown to be intractable (i.e.,...
View Full Document
- Spring '10