Chapter 7 Review

Realistic Scheduling


A realistic schedule includes a detailed knowledge of the work to be done, has task sequences in the correct order, accounts for external constraints, can be accomplished on time, and takes into consideration the objectives of the project.

A predecessor table and a network diagram are two ways of recording sequence constraints. Two rules when graphing task relationships with a network diagram: define task relationships only between work packages, and task relationships should reflect only sequence constraints between work packages, not resource constraints.

Milestones have zero duration, but are useful to mark significant events in the life of a project such as major progress points.

• Project start and finish milestones are useful anchors for the network.
• Milestones can be used to mark input from one party to another (aka external dependencies).
• Milestones can represent significant events that are not already represented by a work package or summary task.

The finish-to-start relationship indicates that one task must be completed before its successor can begin. This is the most common type of relationship. Start-to-start relationships allow the successor task to begin when its predecessor begins. Finish-to-finish tasks can start independently of each other.

Bottom-up estimating builds a cost and schedule estimate from the summation of cost and schedule estimates for each work package. Cost estimates come from three sources: labor estimates, equipment estimates, and materials estimates (materials costs should be estimated from product specifications, not bottom up estimates). These three cost sources can be replaced by fixed-cost bids.

Labor and duration are not always related in an intuitive way,  you need to consider productivity. Adding people to simple tasks always reduces the duration. However, for tasks involving knowledge workers, adding more workers does not always result in greater productivity or a shorter duration of the task. Also people who spend all their time on a project tend to be more productive than people who are spread across multiple projects.

Calculating an initial schedule is key to establishing realistic schedules and meeting them. It provides detailed schedule data for every work package including early start (earliest begin date for a task), early finish (earliest finish date), late start (latest date a task can start without delaying the project), late finish (the latest date a task can finish without delaying the project).  Calculating these dates is a three step process:

1. Forward Pass: works through the network diagram from start to finish to determine the early start and early finish for each task.
2. Backward Pass: works through the network diagram from finish to start to determine the late start and late finish dates for every task.
3. Calculate the float: determine which tasks have schedule flexibility and which define the critical path (the task progression that has zero or negative float and must be completed on schedule to keep the project on schedule). 

Gantt charts are the most common way to display a project schedule. The time scaled network can also be used when it is important to condense the network onto less paper. The completed initial schedule has not yet taken into account  people and equipment limitations which must be assigned and leveled. It is most productive to have consistent, continuous use of the fewest resources possible.

Resource leveling begins with the initial schedule and work package resource requirements then follows a four step process:

1. Forecast the resource requirements throughout the project for the initial schedule
2. Identify resource peaks
3. At each peak, delay noncritical tasks within their float
4. Eliminate the remaining peaks by reevaluating the work package estimates

If the resource leveled plan is unrealistic, the PM should restimate work packages and look at delaying tasks within their float to remove the worst resource peaks and valleys. Failing that the next option is to accept a later project completion date.