Critical Chain Project Management
Comparison of CPM and CCPM results: According to the results we found above, project duration by CPM traditional approach is 21 days and the project duration for the same amount of work by using CCPM is 16 Days. Using CCPM:Project Duration can be reduced by 25-40%.,Resources can be utilized effectively.,Project is fully focused on both critical and non-critical tasks Critical Path Method
Steps in CPM:
There are five steps in CPM:
1.Creating a network showing the task names, durations, and precedence relationships, where the four dates are empty, except for the early start of the first task.
2.Making a Forward Pass through the network, determining the early start and early finish for every task.
3.Making a Backward Pass through the network, determining the late start and late finish for every task. Calculating the float for every task.
5.Finding the paths that consist entirely of tasks with zero float. These paths will go from the very first task to the very last one.
- ES for Early Start
- EF for Early Finish
- LS for Late Start
- LF for Late Finish
Forward Pass:
- The forward pass goes from the initial task (the task with no predecessors) to the final task (the one with no successors), visiting every task in every path and setting the ES and EF dates on the tasks.
- The algorithm is similar to graph theory’s depth-first search, except that the forward pass follows every path from initial to final task, while depth-first search stops when it arrives at a task that it’s already visited. When the forward pass arrives at a task, it may change that task’s ES and EF dates, and that change must be carried forward.
- If the task has already been visited and this visit doesn’t change its dates, there is no need to go forward from this task, since nothing beyond it will change either. During the forward pass, a task may be visited several times as different paths through the network are followed. A task’s ES is determined by the predecessor task with the latest EF, since a task can’t start until all of its predecessors have finished.
The steps of this pass are:
1. Set the initial task’s ES = 0.
2. Calculate the task’s EF = ES duration – 1.
3. Calculate an ES for the task’s successors; ESsuccessor = EFtask 1.
4. Visit all of the task’s successors. If the successor’s current ES is less than the ES from step 3:
a. Replace the current ES with the step 3 value.
b. Repeat steps 2–4 for the current task’s successors, unless there are no more because the last task has been reached.
5. Go back up the path just completed to a task with unvisited successors.
6. Repeat steps 4–5 until all successors have been visited, which will cover all paths from those successors to the final task.
1. Task A: ESA = 0, EFA = ESA dA – 1 = 3
2. Task B: ESB = EFA 1 = 4, EFB = ESB dB – 1 = 8
3. Task E: ESE = EFB 1 = 9, EFE = ESE dE – 1 = 13
4. Task F: ESF = EFE 1 = 14, EFF = ESF dF – 1 = 14
5. Task I: ESI = EFF 1 = 15, EFI = ESI dI – 1 = 19 (Back up to E, next task is G)
6. Task G: ESG = EFE 1 = 14, EFG = ESG dG – 1 = 15
7. Task H: ESH = EFG 1 = 16, EFH = ESH dH – 1 = 18
8. Task I: ESI = EFH 1 = 19, EFI = ESI dI – 1 = 23, replacing step 5’s smaller ES and EF (Back up to E, no next task; back up to A, next task is C)
9. Task C: ESC = EFA 1 = 4, EFC = ESC dC – 1 = 4
10. Task D: ESD = EFC 1 = 5, EFD = ESD dD – 1 = 6
11. Task E: possible new ESE = EFD 1 = 7, EFE = ESE dE – 1 = 11, smaller than step 3’s ES and EF, no replacement (no change to propagate forward) (Back up to A, no next task)
12. The forward pass stops here, since there are no more paths to probe. Notice the multiple visits to several tasks over different paths.